Current Text-to-Code models demonstrate impressive capabilities in generating executable code from natural language snippets. However, current studies focus on technical instructions and programmer-oriented language, and it is an open question whether these models can effectively translate natural language descriptions given by non-technical users and express complex goals, to an executable program that contains an intricate flow—composed of API access and control structures as loops, conditions, and sequences. To unlock the challenge of generating a complete program from a plain non-technical description we present NoviCode, a novel NL Programming task, which takes as input an API and a natural language description by a novice non-programmer, and provides an executable program as output. To assess the efficacy of models on this task, we provide a novel benchmark accompanied by test suites wherein the generated program code is assessed not according to their form, but according to their functional execution. Our experiments show that, first, NoviCode is indeed a challenging task in the code synthesis domain, and that generating complex code from non-technical instructions goes beyond the current Text-to-Code paradigm. Second, we show that a novel approach wherein we align the NL utterances with the compositional hierarchical structure of the code, greatly enhances the performance of LLMs on this task, compared with the end-to-end Text-to-Code counterparts.

Document collections of various domains, e.g., legal, medical, or financial, often share some underlying collection-wide structure, which captures information that can aid both human users and structure-aware models.We propose to identify the typical structure of document within a collection, which requires to capture recurring topics across the collection, while abstracting over arbitrary header paraphrases, and ground each topic to respective document locations. These requirements pose several challenges: headers that mark recurring topics frequently differ in phrasing, certain section headers are unique to individual documents and do not reflect the typical structure, and the order of topics can vary between documents. Subsequently, we develop an unsupervised graph-based method which leverages both inter- and intra-document similarities, to extract the underlying collection-wide structure. Our evaluations on three diverse domains in both English and Hebrew indicate that our method extracts meaningful collection-wide structure, and we hope that future work will leverage our method for multi-document applications and structure-aware models.

The exponential growth of scientific literature necessitates advanced tools for effective knowledge exploration. We present Knowledge Navigator, a system designed to enhance exploratory search abilities by organizing and structuring the retrieved documents from broad topical queries into a navigable, two-level hierarchy of named and descriptive scientific topics and subtopics. This structured organization provides an overall view of the research themes in a domain, while also enabling iterative search and deeper knowledge discovery within specific subtopics by allowing users to refine their focus and retrieve additional relevant documents. Knowledge Navigator combines LLM capabilities with cluster-based methods to enable an effective browsing method. We demonstrate our approach’s effectiveness through automatic and manual evaluations on two novel benchmarks, CLUSTREC-COVID and SCITOC Our code, prompts, and benchmarks are made publicly available.

While coreference resolution is traditionally used as a component in individual document understanding, in this work we take a more global view and explore what can we learn about a domain from the set of all document-level coreference relations that are present in a large corpus. We derive coreference chains from a corpus of 30 million biomedical abstracts and construct a graph based on the string phrases within these chains, establishing connections between phrases if they co-occur within the same coreference chain. We then use the graph structure and the betweeness centrality measure to distinguish between edges denoting hierarchy, identity and noise, assign directionality to edges denoting hierarchy, and split nodes (strings) that correspond to multiple distinct concepts. The result is a rich, data-driven ontology over concepts in the biomedical domain, parts of which overlaps significantly with human-authored ontologies. We release the coreference chains and resulting ontology under a creative-commons license.

The escalating debate on AI’s capabilities warrants developing reliable metrics to assess machine “intelligence.” Recently, many anecdotal examples were used to suggest that newer Large Language Models (LLMs) like ChatGPT and GPT-4 exhibit Neural Theory-of-Mind (N-ToM); however, prior work reached conflicting conclusions regarding those abilities. We investigate the extent of LLMs’ N-ToM through an extensive evaluation of 6 tasks and find that while LLMs exhibit certain N-ToM abilities, this behavior is far from being robust. We further examine the factors impacting performance on N-ToM tasks and discover that LLMs struggle with adversarial examples, indicating reliance on shallow heuristics rather than robust ToM abilities. We caution against drawing conclusions from anecdotal examples, limited benchmark testing, and using human-designed psychological tests to evaluate models.

The 1st Workshop on Data Contamination (CONDA 2024) focuses on all relevant aspects of data contamination in natural language processing, where data contamination is understood as situations where evaluation data is included in pre-training corpora used to train large scale models, compromising evaluation results. The workshop fostered a shared task to collect evidence on data contamination in current available datasets and models. The goal of the shared task and associated database is to assist the community in understanding the extent of the problem and to assist researchers in avoiding reporting evaluation results on known contaminated resources. The shared task provides a structured, centralized public database for the collection of contamination evidence, open to contributions from the community via GitHub pool requests. This first compilation paper is based on 566 reported entries over 91 contaminated sources from a total of 23 contributors. The details of the individual contamination events are available in the platform. The platform continues to be online, open to contributions from the community.

This paper explores the impact of extending input lengths on the capabilities of Large Language Models (LLMs). Despite LLMs advancements in recent times, their performance consistency across different input lengths is not well understood. We investigate this aspect by introducing a novel QA reasoning framework, specifically designed to assess the impact of input length. We isolate the effect of input length using multiple versions of the same sample, each being extended with padding of different lengths, types and locations. Our findings show a notable degradation in LLMs’ reasoning performance at much shorter input lengths than their technical maximum. We show that the degradation trend appears in every version of our dataset, although at different intensities.Additionally, our study reveals that the traditional metric of next word prediction correlates negatively with performance of LLMs’ on our reasoning dataset. We analyse our results and identify failure modes that can serve as useful guides for future research, potentially informing strategies to address the limitations observed in LLMs.

Methods for erasing human-interpretable concepts from neural representations that assume linearity have been found to be tractable and useful. However, the impact of this removal on the behavior of downstream classifiers trained on the modified representations is not fully understood. In this work, we formally define the notion of linear guardedness as the inability of an adversary to predict the concept directly from the representation, and study its implications. We show that, in the binary case, under certain assumptions, a downstream log-linear model cannot recover the erased concept. However, we constructively demonstrate that a multiclass log-linear model can be constructed that indirectly recovers the concept in some cases, pointing to the inherent limitations of linear guardedness as a downstream bias mitigation technique.These findings shed light on the theoretical limitations of linear erasure methods and highlight the need for further research on the connections between intrinsic and extrinsic bias in neural models.

Verbal omissions are complex syntactic phenomena in VP coordination structures. They occur when verbs and (some of) their arguments are omitted from subsequent clauses after being explicitly stated in an initial clause. Recovering these omitted elements is necessary for accurate interpretation of the sentence, and while humans easily and intuitively fill in the missing information, state-of-the-art models continue to struggle with this task. Previous work is limited to small-scale datasets, synthetic data creation methods, and to resolution methods in the dependency-graph level. In this work we propose a conjunct resolution task that operates directly on the text and makes use of a split-and-rephrase paradigm in order to recover the missing elements in the coordination structure. To this end, we first formulate a pragmatic framework of verbal omissions which describes the different types of omissions, and develop an automatic scalable collection method. Based on this method, we curate a large dataset, containing over 10K examples of naturally-occurring verbal omissions with crowd-sourced annotations of the resolved conjuncts. We train various neural baselines for this task, and show that while our best method obtains decent performance, it leaves ample space for improvement. We propose our dataset, metrics and models as a starting point for future research on this topic.

Information extraction systems often producehundreds to thousands of strings on a specifictopic. We present a method that facilitatesbetter consumption of these strings, in an ex-ploratory setting in which a user wants to bothget a broad overview of what’s available, and achance to dive deeper on some aspects. The sys-tem works by grouping similar items together,and arranging the remaining items into a hierar-chical navigable DAG structure. We apply themethod to medical information extraction.

We propose a new commonsense reasoning benchmark to motivate commonsense reasoning progress from two perspectives: (1) Evaluating whether models can distinguish knowledge quality by predicting if the knowledge is enough to answer the question; (2) Evaluating whether models can develop commonsense inference capabilities that generalize across tasks. We first extract supporting knowledge for each question and ask humans to annotate whether the auto-extracted knowledge is enough to answer the question or not. After that, we convert different tasks into a unified question-answering format to evaluate the models’ generalization capabilities. We name the benchmark Commonsense Inference with Knowledge-in-the-loop Question Answering (\name). Experiments show that with our learning paradigm, models demonstrate encouraging generalization capabilities. At the same time, we also notice that distinguishing knowledge quality remains challenging for current commonsense reasoning models.

Pre-trained neural masked language models are often used for predicting a replacement token for a given sequence position, in a cloze-like task. However, this usage is restricted to predicting a single token, from a relatively small pre-trained vocabulary. Recent Sequence2Sequence pre-trained LMs like T5 do allow predicting multi-token completions, but are more expensive to train and run. We show that pre-trained masked language models can be adapted to produce multi-token completions, with only a modest addition to their parameter count. We propose two simple adaptation approaches, trading parameter counts for accuracy. The first method generates multi-token completions from a conditioned RNN. It has a very low parameter count and achieves competitive results. The second method is even simpler: it adds items corresponding to multi-token units to the output prediction matrix. While being higher in parameter count than the RNN method, it also surpasses current state-of-the-art multi-token completion models, including T5-3B, while being significantly more parameter efficient. We demonstrate that our approach is flexible to different vocabularies and domains and can effectively leverage existing pre-trained models available in different domains. Finally, a human evaluation further validates our results and shows that our solution regularly provides valid completions, as well as reasonable correctness for factual-sentence completions.

Language models generate text based on successively sampling the next word. A decoding procedure based on nucleus (top-p) sampling chooses from the smallest possible set of words whose cumulative probability exceeds the probability p. In this work, we assess whether a top-p set is indeed aligned with its probabilistic meaning in various linguistic contexts.We employ conformal prediction, a calibration procedure that focuses on the construction of minimal prediction sets according to a desired confidence level, to calibrate the parameter p as a function of the entropy of the next word distribution. We find that OPT models are overconfident, and that calibration shows a moderate inverse scaling with model size.

Deverbal nouns are nominal forms of verbs commonly used in written English texts to describe events or actions, as well as their arguments. However, many NLP systems, and in particular pattern-based ones, neglect to handle such nominalized constructions. The solutions that do exist for handling arguments of nominalized constructions are based on semantic annotation and require semantic ontologies, making their applications restricted to a small set of nouns. We propose to adopt instead a more syntactic approach, which maps the arguments of deverbal nouns to the universal-dependency relations of the corresponding verbal construction. We present an unsupervised mechanism—based on contextualized word representations—which allows to enrich universal-dependency trees with dependency arcs denoting arguments of deverbal nouns, using the same labels as the corresponding verbal cases. By sharing the same label set as in the verbal case, patterns that were developed for verbs can be applied without modification but with high accuracy also to the nominal constructions.

Motivated by the question of the extent to which large language models “understand” social intelligence, we investigate the ability of such models to generate correct responses to questions involving descriptions of faux pas situations. The faux pas test is a test used in clinical psychology, which is known to be more challenging for children than individual tests of theory-of-mind or social intelligence. Our results demonstrate that, while the models seem to sometimes offer correct responses, they in fact struggle with this task, and that many of the seemingly correct responses can be attributed to over-interpretation by the human reader (“the ELIZA effect”). An additional phenomenon observed is the failure of most models to generate a correct response to presupposition questions. Finally, in an experiment in which the models are tasked with generating original faux pas stories, we find that while some models are capable of generating novel faux pas stories, the stories are all explicit, as the models are limited in their abilities to describe situations in an implicit manner.

Word-level saliency explanations (“heat maps over words”) are often used to communicate feature-attribution in text-based models. Recent studies found that superficial factors such as word length can distort human interpretation of the communicated saliency scores. We conduct a user study to investigate how the marking of a word’s *neighboring words* affect the explainee’s perception of the word’s importance in the context of a saliency explanation. We find that neighboring words have significant effects on the word’s importance rating. Concretely, we identify that the influence changes based on neighboring direction (left vs. right) and a-priori linguistic and computational measures of phrases and collocations (vs. unrelated neighboring words).Our results question whether text-based saliency explanations should be continued to be communicated at word level, and inform future research on alternative saliency explanation methods.

Recognizing entities in texts is a central need in many information-seeking scenarios, and indeed, Named Entity Recognition (NER) is arguably one of the most successful examples of a widely adopted NLP task and corresponding NLP technology. Recent advances in large language models (LLMs) appear to provide effective solutions (also) for NER tasks that were traditionally handled with dedicated models, often matching or surpassing the abilities of the dedicated models. Should NER be considered a solved problem? We argue to the contrary: the capabilities provided by LLMs are not the end of NER research, but rather an exciting beginning. They allow taking NER to the next level, tackling increasingly more useful, and increasingly more challenging, variants. We present three variants of the NER task, together with a dataset to support them. The first is a move towards more fine-grained—and intersectional—entity types. The second is a move towards zero-shot recognition and extraction of these fine-grained types based on entity-type labels. The third, and most challenging, is the move from the recognition setup to a novel retrieval setup, where the query is a zero-shot entity type, and the expected result is all the sentences from a large, pre-indexed corpus that contain entities of these types, and their corresponding spans. We show that all of these are far from being solved. We provide a large, silver-annotated corpus of 4 million paragraphs covering 500 entity types, to facilitate research towards all of these three goals.

Recent applications of LLMs in Machine Reading Comprehension (MRC) systems have shown impressive results, but the use of shortcuts, mechanisms triggered by features spuriously correlated to the true label, has emerged as a potential threat to their reliability. We analyze the problem from two angles: LLMs as editors, guided to edit text to mislead LLMs; and LLMs as readers, who answer questions based on the edited text. We introduce a framework that guides an editor to add potential shortcuts-triggers to samples. Using GPT4 as the editor, we find it can successfully edit trigger shortcut in samples that fool LLMs. Analysing LLMs as readers, we observe that even capable LLMs can be deceived using shortcut knowledge. Strikingly, we discover that GPT4 can be deceived by its own edits (15% drop in F1). Our findings highlight inherent vulnerabilities of LLMs to shortcut manipulations. We publish ShortcutQA, a curated dataset generated by our framework for future research.

Sign language segmentation is a crucial task in sign language processing systems. It enables downstream tasks such as sign recognition, transcription, and machine translation. In this work, we consider two kinds of segmentation: segmentation into individual signs and segmentation into phrases, larger units comprising several signs. We propose a novel approach to jointly model these two tasks. Our method is motivated by linguistic cues observed in sign language corpora. We replace the predominant IO tagging scheme with BIO tagging to account for continuous signing. Given that prosody plays a significant role in phrase boundaries, we explore the use of optical flow features. We also provide an extensive analysis of hand shapes and 3D hand normalization. We find that introducing BIO tagging is necessary to model sign boundaries. Explicitly encoding prosody by optical flow improves segmentation in shallow models, but its contribution is negligible in deeper models. Careful tuning of the decoding algorithm atop the models further improves the segmentation quality. We demonstrate that our final models generalize to out-of-domain video content in a different signed language, even under a zero-shot setting. We observe that including optical flow and 3D hand normalization enhances the robustness of the model in this context.

Current benchmarks for Hebrew Natural Language Processing (NLP) focus mainly on morpho-syntactic tasks, neglecting the semantic dimension of language understanding. To bridge this gap, we set out to deliver a Hebrew Machine Reading Comprehension (MRC) dataset, where MRC is to be realized as extractive Question Answering. The morphologically-rich nature of Hebrew poses a challenge to this endeavor: the indeterminacy and non-transparency of span boundaries in morphologically complex forms lead to annotation inconsistencies, disagreements, and flaws of standard evaluation metrics. To remedy this, we devise a novel set of guidelines, a controlled crowdsourcing protocol, and revised evaluation metrics, that are suitable for the morphologically rich nature of the language. Our resulting benchmark, HeQ (Hebrew QA), features 30,147 diverse question-answer pairs derived from both Hebrew Wikipedia articles and Israeli tech news. Our empirical investigation reveals that standard evaluation metrics such as F1 Scores and Exact Match (EM) are not appropriate for Hebrew (and other MRLs), and we propose a relevant enhancement. In addition, our experiments show low correlation between models’ performance on morpho-syntactic tasks and on MRC, which suggests that models that are designed for the former might underperform on semantic-heavy tasks. The development and exploration of HeQ illustrate some of the challenges MRLs pose in natural language understanding (NLU), fostering progression towards more and better NLU models for Hebrew and other MRLs.

Data contamination has become prevalent and challenging with the rise of models pretrained on large automatically-crawled corpora. For closed models, the training data becomes a trade secret, and even for open models, it is not trivial to detect contamination. Strategies such as leaderboards with hidden answers, or using test data which is guaranteed to be unseen, are expensive and become fragile with time. Assuming that all relevant actors value clean test data and will cooperate to mitigate data contamination, what can be done? We propose three strategies that can make a difference: (1) Test data made public should be encrypted with a public key and licensed to disallow derivative distribution; (2) demand training exclusion controls from closed API holders, and protect your test data by refusing to evaluate without them; (3) avoid data which appears with its solution on the internet, and release the web-page context of internet-derived data along with the data. These strategies are practical and can be effective in preventing data contamination.

To produce accurate predictions, language models (LMs) must balance between generalization and memorization. Yet, little is known about the mechanism by which transformer LMs employ their memorization capacity. When does a model decide to output a memorized phrase, and how is this phrase then retrieved from memory? In this work, we offer the first methodological framework for probing and characterizing recall of memorized sequences in transformer LMs. First, we lay out criteria for detecting model inputs that trigger memory recall, and propose idioms as inputs that typically fulfill these criteria. Next, we construct a dataset of English idioms and use it to compare model behavior on memorized vs. non-memorized inputs. Specifically, we analyze the internal prediction construction process by interpreting the model’s hidden representations as a gradual refinement of the output probability distribution. We find that across different model sizes and architectures, memorized predictions are a two-step process: early layers promote the predicted token to the top of the output distribution, and upper layers increase model confidence. This suggests that memorized information is stored and retrieved in the early layers of the network. Last, we demonstrate the utility of our methodology beyond idioms in memorized factual statements. Overall, our work makes a first step towards understanding memory recall, and provides a methodological basis for future studies of transformer memorization.

Current state-of-the-art coreference systems are based on a single pairwise scoring component, which assigns to each pair of mention spans a score reflecting their tendency to corefer to each other. We observe that different kinds of mention pairs require different information sources to assess their score. We present LingMess, a linguistically motivated categorization of mention-pairs into 6 types of coreference decisions and learn a dedicated trainable scoring function for each category. This significantly improves the accuracy of the pairwise scorer as well as of the overall coreference performance on the English Ontonotes coreference corpus and 5 additional datasets.

Sign language translation systems are complex and require many components. As a result, it is very hard to compare methods across publications. We present an open-source implementation of a text-to-gloss-to-pose-to-video pipeline approach, demonstrating conversion from German to Swiss German Sign Language, French to French Sign Language of Switzerland, and Italian to Italian Sign Language of Switzerland. We propose three different components for the text-to-gloss translation: a lemmatizer, a rule-based word reordering and dropping component, and a neural machine translation system. Gloss-to-pose conversion occurs using data from a lexicon for three different signed languages, with skeletal poses extracted from videos. To generate a sentence, the text-to-gloss system is first run, and the pose representations of the resulting signs are stitched together.

Combination therapies have become the standard of care for diseases such as cancer, tuberculosis, malaria and HIV. However, the combinatorial set of available multi-drug treatments creates a challenge in identifying effective combination therapies available in a situation. To assist medical professionals in identifying beneficial drug-combinations, we construct an expert-annotated dataset for extracting information about the efficacy of drug combinations from the scientific literature. Beyond its practical utility, the dataset also presents a unique NLP challenge, as the first relation extraction dataset consisting of variable-length relations. Furthermore, the relations in this dataset predominantly require language understanding beyond the sentence level, adding to the challenge of this task. We provide a promising baseline model and identify clear areas for further improvement. We release our dataset (https://huggingface.co/datasets/allenai/drug-combo-extraction), code (https://github.com/allenai/drug-combo-extraction) and baseline models (https://huggingface.co/allenai/drug-combo-classifier-pubmedbert-dapt) publicly to encourage the NLP community to participate in this task.

Multilingual language models were shown to allow for nontrivial transfer across scripts and languages. In this work, we study the structure of the internal representations that enable this transfer. We focus on the representations of gender distinctions as a practical case study, and examine the extent to which the gender concept is encoded in shared subspaces across different languages. Our analysis shows that gender representations consist of several prominent components that are shared across languages, alongside language-specific components. The existence of language-independent and language-specific components provides an explanation for an intriguing empirical observation we make”:” while gender classification transfers well across languages, interventions for gender removal trained on a single language do not transfer easily to others.

We study the way DALLE-2 maps symbols (words) in the prompt to their references (entities or properties of entities in the generated image). We show that in stark contrast to the way human process language, DALLE-2 does not follow the constraint that each word has a single role in the interpretation, and sometimes re-use the same symbol for different purposes. We collect a set of stimuli that reflect the phenomenon: we show that DALLE-2 depicts both senses of nouns with multiple senses at once; and that a given word can modify the properties of two distinct entities in the image, or can be depicted as one object and also modify the properties of another object, creating a semantic leakage of properties between entities. Taken together, our study highlights the differences between DALLE-2 and human language processing and opens an avenue for future study on the inductive biases of text-to-image models.

We present a word-sense induction method based on pre-trained masked language models (MLMs), which can cheaply scale to large vocabularies and large corpora. The result is a corpus which is sense-tagged according to a corpus-derived sense inventory and where each sense is associated with indicative words. Evaluation on English Wikipedia that was sense-tagged using our method shows that both the induced senses, and the per-instance sense assignment, are of high quality even compared to WSD methods, such as Babelfy. Furthermore, by training a static word embeddings algorithm on the sense-tagged corpus, we obtain high-quality static senseful embeddings. These outperform existing senseful embeddings methods on the WiC dataset and on a new outlier detection dataset we developed. The data driven nature of the algorithm allows to induce corpora-specific senses, which may not appear in standard sense inventories, as we demonstrate using a case study on the scientific domain.

We introduce BitFit, a sparse-finetuning method where only the bias-terms of the model (or a subset of them) are being modified. We show that with small-to-medium training data, applying BitFit on pre-trained BERT models is competitive with (and sometimes better than) fine-tuning the entire model. For larger data, the method is competitive with other sparse fine-tuning methods. Besides their practical utility, these findings are relevant for the question of understanding the commonly-used process of finetuning: they support the hypothesis that finetuning is mainly about exposing knowledge induced by language-modeling training, rather than learning new task-specific linguistic knowledge.

Transformer-based language models (LMs) are at the core of modern NLP, but their internal prediction construction process is opaque and largely not understood. In this work, we make a substantial step towards unveiling this underlying prediction process, by reverse-engineering the operation of the feed-forward network (FFN) layers, one of the building blocks of transformer models. We view the token representation as a changing distribution over the vocabulary, and the output from each FFN layer as an additive update to that distribution. Then, we analyze the FFN updates in the vocabulary space, showing that each update can be decomposed to sub-updates corresponding to single FFN parameter vectors, each promoting concepts that are often human-interpretable. We then leverage these findings for controlling LM predictions, where we reduce the toxicity of GPT2 by almost 50%, and for improving computation efficiency with a simple early exit rule, saving 20% of computation on average.

The representation space of neural models for textual data emerges in an unsupervised manner during training. Understanding how human-interpretable concepts, such as gender, are encoded in these representations would improve the ability of users to control the content of these representations and analyze the working of the models that rely on them. One prominent approach to the control problem is the identification and removal of linear concept subspaces – subspaces in the representation space that correspond to a given concept. While those are tractable and interpretable, neural network do not necessarily represent concepts in linear subspaces. We propose a kernelization of the recently-proposed linear concept-removal objective, and show that it is effective in guarding against the ability of certain nonlinear adversaries to recover the concept. Interestingly, our findings suggest that the division between linear and nonlinear models is overly simplistic: when considering the concept of binary gender and its neutralization, we do not find a single kernel space that exclusively contains all the concept-related information. It is therefore challenging to protect against all nonlinear adversaries at once.

The opaque nature and unexplained behavior of transformer-based language models (LMs) have spurred a wide interest in interpreting their predictions. However, current interpretation methods mostly focus on probing models from outside, executing behavioral tests, and analyzing salience input features, while the internal prediction construction process is largely not understood. In this work, we introduce LM-Debugger, an interactive debugger tool for transformer-based LMs, which provides a fine-grained interpretation of the model’s internal prediction process, as well as a powerful framework for intervening in LM behavior. For its backbone, LM-Debugger relies on a recent method that interprets the inner token representations and their updates by the feed-forward layers in the vocabulary space. We demonstrate the utility of LM-Debugger for single-prediction debugging, by inspecting the internal disambiguation process done by GPT2. Moreover, we show how easily LM-Debugger allows to shift model behavior in a direction of the user’s choice, by identifying a few vectors in the network and inducing effective interventions to the prediction process. We release LM-Debugger as an open-source tool and a demo over GPT2 models.

Phrase similarity is a key component of many NLP applications. Current phrase similarity methods focus on embedding the phrase itself and use the phrase context only during training of the pretrained model. To better leverage the information in the context, we propose McPhraSy (Multi-context Phrase Similarity), a novel algorithm for estimating the similarity of phrases based on multiple contexts. At inference time, McPhraSy represents each phrase by considering multiple contexts in which it appears and computes the similarity of two phrases by aggregating the pairwise similarities between the contexts of the phrases. Incorporating context during inference enables McPhraSy to outperform current state-of-the-art models on two phrase similarity datasets by up to 13.3%. Finally, we also present a new downstream task that relies on phrase similarity – keyphrase clustering – and create a new benchmark for it in the product reviews domain. We show that McPhraSy surpasses all other baselines for this task.

While fine-tuned language models perform well on many language tasks, they were also shown to rely on superficial surface features such as lexical overlap. Excessive utilization of such heuristics can lead to failure on challenging inputs. We analyze the use of lexical overlap heuristics in natural language inference, paraphrase detection, and reading comprehension (using a novel contrastive dataset),and find that larger models are much less susceptible to adopting lexical overlap heuristics. We also find that longer training leads models to abandon lexical overlap heuristics. Finally, We provide evidence that the disparity between models size has its source in the pre-trained model.

We introduce fastcoref, a python package for fast, accurate, and easy-to-use English coreference resolution. The package is pip-installable, and allows two modes: an accurate mode based on the LingMess architecture, providing state-of-the-art coreference accuracy, and a substantially faster model, F-coref, which is the focus of this work. F-coref allows to process 2.8K OntoNotes documents in 25 seconds on a V100 GPU (compared to 6 minutes for the LingMess model, and to 12 minutes of the popular AllenNLP coreference model) with only a modest drop in accuracy. The fast speed is achieved through a combination of distillation of a compact model from the LingMess model, and an efficient batching implementation using a technique we call leftover batching. https://github.com/shon-otmazgin/fastcoref

Understanding the relations between entities denoted by NPs in a text is a critical part of human-like natural language understanding. However, only a fraction of such relations is covered by standard NLP tasks and benchmarks nowadays. In this work, we propose a novel task termed text-based NP enrichment (TNE), in which we aim to enrich each NP in a text with all the preposition-mediated relations—either explicit or implicit—that hold between it and other NPs in the text. The relations are represented as triplets, each denoted by two NPs related via a preposition. Humans recover such relations seamlessly, while current state-of-the-art models struggle with them due to the implicit nature of the problem. We build the first large-scale dataset for the problem, provide the formal framing and scope of annotation, analyze the data, and report the results of fine-tuned language models on the task, demonstrating the challenge it poses to current technology. A webpage with a data-exploration UI, a demo, and links to the code, models, and leaderboard, to foster further research into this challenging problem can be found at: yanaiela.github.io/TNE/.

Signed languages are the primary means of communication for many deaf and hard of hearing individuals. Since signed languages exhibit all the fundamental linguistic properties of natural language, we believe that tools and theories of Natural Language Processing (NLP) are crucial towards its modeling. However, existing research in Sign Language Processing (SLP) seldom attempt to explore and leverage the linguistic organization of signed languages. This position paper calls on the NLP community to include signed languages as a research area with high social and scientific impact. We first discuss the linguistic properties of signed languages to consider during their modeling. Then, we review the limitations of current SLP models and identify the open challenges to extend NLP to signed languages. Finally, we urge (1) the adoption of an efficient tokenization method; (2) the development of linguistically-informed models; (3) the collection of real-world signed language data; (4) the inclusion of local signed language communities as an active and leading voice in the direction of research.

Domain experts often need to extract structured information from large corpora. We advocate for a search paradigm called “extractive search”, in which a search query is enriched with capture-slots, to allow for such rapid extraction. Such an extractive search system can be built around syntactic structures, resulting in high-precision, low-recall results. We show how the recall can be improved using neural retrieval and alignment. The goals of this paper are to concisely introduce the extractive-search paradigm; and to demonstrate a prototype neural retrieval system for extractive search and its benefits and potential. Our prototype is available at https://spike.neural-sim.apps.allenai.org/ and a video demonstration is available at https://vimeo.com/559586687.

A growing body of work makes use of probing in order to investigate the working of neural models, often considered black boxes. Recently, an ongoing debate emerged surrounding the limitations of the probing paradigm. In this work, we point out the inability to infer behavioral conclusions from probing results, and offer an alternative method that focuses on how the information is being used, rather than on what information is encoded. Our method, Amnesic Probing, follows the intuition that the utility of a property for a given task can be assessed by measuring the influence of a causal intervention that removes it from the representation. Equipped with this new analysis tool, we can ask questions that were not possible before, for example, is part-of-speech information important for word prediction? We perform a series of analyses on BERT to answer these types of questions. Our findings demonstrate that conventional probing performance is not correlated to task importance, and we call for increased scrutiny of claims that draw behavioral or causal conclusions from probing results.1

We find that the requirement of model interpretations to be faithful is vague and incomplete. With interpretation by textual highlights as a case study, we present several failure cases. Borrowing concepts from social science, we identify that the problem is a misalignment between the causal chain of decisions (causal attribution) and the attribution of human behavior to the interpretation (social attribution). We reformulate faithfulness as an accurate attribution of causality to the model, and introduce the concept of aligned faithfulness: faithful causal chains that are aligned with their expected social behavior. The two steps of causal attribution and social attribution together complete the process of explaining behavior. With this formalization, we characterize various failures of misaligned faithful highlight interpretations, and propose an alternative causal chain to remedy the issues. Finally, we implement highlight explanations of the proposed causal format using contrastive explanations.

We explore few-shot learning (FSL) for relation classification (RC). Focusing on the realistic scenario of FSL, in which a test instance might not belong to any of the target categories (none-of-the-above, [NOTA]), we first revisit the recent popular dataset structure for FSL, pointing out its unrealistic data distribution. To remedy this, we propose a novel methodology for deriving more realistic few-shot test data from available datasets for supervised RC, and apply it to the TACRED dataset. This yields a new challenging benchmark for FSL-RC, on which state of the art models show poor performance. Next, we analyze classification schemes within the popular embedding-based nearest-neighbor approach for FSL, with respect to constraints they impose on the embedding space. Triggered by this analysis, we propose a novel classification scheme in which the NOTA category is represented as learned vectors, shown empirically to be an appealing option for FSL.

Consistency of a model—that is, the invariance of its behavior under meaning-preserving alternations in its input—is a highly desirable property in natural language processing. In this paper we study the question: Are Pretrained Language Models (PLMs) consistent with respect to factual knowledge? To this end, we create ParaRel🤘, a high-quality resource of cloze-style query English paraphrases. It contains a total of 328 paraphrases for 38 relations. Using ParaRel🤘, we show that the consistency of all PLMs we experiment with is poor— though with high variance between relations. Our analysis of the representational spaces of PLMs suggests that they have a poor structure and are currently not suitable for representing knowledge robustly. Finally, we propose a method for improving model consistency and experimentally demonstrate its effectiveness.1

Language models trained on billions of tokens have recently led to unprecedented results on many NLP tasks. This success raises the question of whether, in principle, a system can ever “understand” raw text without access to some form of grounding. We formally investigate the abilities of ungrounded systems to acquire meaning. Our analysis focuses on the role of “assertions”: textual contexts that provide indirect clues about the underlying semantics. We study whether assertions enable a system to emulate representations preserving semantic relations like equivalence. We find that assertions enable semantic emulation of languages that satisfy a strong notion of semantic transparency. However, for classes of languages where the same expression can take different values in different contexts, we show that emulation can become uncomputable. Finally, we discuss differences between our formal model and natural language, exploring how our results generalize to a modal setting and other semantic relations. Together, our results suggest that assertions in code or language do not provide sufficient signal to fully emulate semantic representations. We formalize ways in which ungrounded language models appear to be fundamentally limited in their ability to “understand”.

During production of this paper, an error was introduced to the formula on the bottom of the right column of page 1020. In the last two terms of the formula, the n and m subscripts were swapped. The correct formula is:Lc=∑n=1k∑m=n+1kDKL(Qnri∥Qmri)+DKL(Qmri∥Qnri)The paper has been updated.

Sign language translation (SLT) is often decomposed into video-to-gloss recognition and gloss to-text translation, where a gloss is a sequence of transcribed spoken-language words in the order in which they are signed. We focus here on gloss-to-text translation, which we treat as a low-resource neural machine translation (NMT) problem. However, unlike traditional low resource NMT, gloss-to-text translation differs because gloss-text pairs often have a higher lexical overlap and lower syntactic overlap than pairs of spoken languages. We exploit this lexical overlap and handle syntactic divergence by proposing two rule-based heuristics that generate pseudo-parallel gloss-text pairs from monolingual spoken language text. By pre-training on this synthetic data, we improve translation from American Sign Language (ASL) to English and German Sign Language (DGS) to German by up to 3.14 and 2.20 BLEU, respectively.

Large Transformers pretrained over clinical notes from Electronic Health Records (EHR) have afforded substantial gains in performance on predictive clinical tasks. The cost of training such models (and the necessity of data access to do so) coupled with their utility motivates parameter sharing, i.e., the release of pretrained models such as ClinicalBERT. While most efforts have used deidentified EHR, many researchers have access to large sets of sensitive, non-deidentified EHR with which they might train a BERT model (or similar). Would it be safe to release the weights of such a model if they did? In this work, we design a battery of approaches intended to recover Personal Health Information (PHI) from a trained BERT. Specifically, we attempt to recover patient names and conditions with which they are associated. We find that simple probing methods are not able to meaningfully extract sensitive information from BERT trained over the MIMIC-III corpus of EHR. However, more sophisticated “attacks” may succeed in doing so: To facilitate such research, we make our experimental setup and baseline probing models available at https://github.com/elehman16/exposing_patient_data_release.

Historical linguists have identified regularities in the process of historic sound change. The comparative method utilizes those regularities to reconstruct proto-words based on observed forms in daughter languages. Can this process be efficiently automated? We address the task of proto-word reconstruction, in which the model is exposed to cognates in contemporary daughter languages, and has to predict the proto word in the ancestor language. We provide a novel dataset for this task, encompassing over 8,000 comparative entries, and show that neural sequence models outperform conventional methods applied to this task so far. Error analysis reveals a variability in the ability of neural model to capture different phonological changes, correlating with the complexity of the changes. Analysis of learned embeddings reveals the models learn phonologically meaningful generalizations, corresponding to well-attested phonological shifts documented by historical linguistics.

When language models process syntactically complex sentences, do they use their representations of syntax in a manner that is consistent with the grammar of the language? We propose AlterRep, an intervention-based method to address this question. For any linguistic feature of a given sentence, AlterRep generates counterfactual representations by altering how the feature is encoded, while leaving in- tact all other aspects of the original representation. By measuring the change in a model’s word prediction behavior when these counterfactual representations are substituted for the original ones, we can draw conclusions about the causal effect of the linguistic feature in question on the model’s behavior. We apply this method to study how BERT models of different sizes process relative clauses (RCs). We find that BERT variants use RC boundary information during word prediction in a manner that is consistent with the rules of English grammar; this RC boundary information generalizes to a considerable extent across different RC types, suggesting that BERT represents RCs as an abstract linguistic category.

We consider the situation in which a user has collected a small set of documents on a cohesive topic, and they want to retrieve additional documents on this topic from a large collection. Information Retrieval (IR) solutions treat the document set as a query, and look for similar documents in the collection. We propose to extend the IR approach by treating the problem as an instance of positive-unlabeled (PU) learning—i.e., learning binary classifiers from only positive (the query documents) and unlabeled (the results of the IR engine) data. Utilizing PU learning for text with big neural networks is a largely unexplored field. We discuss various challenges in applying PU learning to the setting, showing that the standard implementations of state-of-the-art PU solutions fail. We propose solutions for each of the challenges and empirically validate them with ablation tests. We demonstrate the effectiveness of the new method using a series of experiments of retrieving PubMed abstracts adhering to fine-grained topics, showing improvements over the common IR solution and other baselines.

The advent of neural-networks in NLP brought with it substantial improvements in supervised relation extraction. However, obtaining a sufficient quantity of training data remains a key challenge. In this work we propose a process for bootstrapping training datasets which can be performed quickly by non-NLP-experts. We take advantage of search engines over syntactic-graphs (Such as Shlain et al. (2020)) which expose a friendly by-example syntax. We use these to obtain positive examples by searching for sentences that are syntactically similar to user input examples. We apply this technique to relations from TACRED and DocRED and show that the resulting models are competitive with models trained on manually annotated data and on data obtained from distant supervision. The models also outperform models trained using NLG data augmentation techniques. Extending the search-based approach with the NLG method further improves the results.

Asking questions about a situation is an inherent step towards understanding it. To this end, we introduce the task of role question generation, which, given a predicate mention and a passage, requires producing a set of questions asking about all possible semantic roles of the predicate. We develop a two-stage model for this task, which first produces a context-independent question prototype for each role and then revises it to be contextually appropriate for the passage. Unlike most existing approaches to question generation, our approach does not require conditioning on existing answers in the text. Instead, we condition on the type of information to inquire about, regardless of whether the answer appears explicitly in the text, could be inferred from it, or should be sought elsewhere. Our evaluation demonstrates that we generate diverse and well-formed questions for a large, broad-coverage ontology of predicates and roles.

Contrastive explanations clarify why an event occurred in contrast to another. They are inherently intuitive to humans to both produce and comprehend. We propose a method to produce contrastive explanations in the latent space, via a projection of the input representation, such that only the features that differentiate two potential decisions are captured. Our modification allows model behavior to consider only contrastive reasoning, and uncover which aspects of the input are useful for and against particular decisions. Our contrastive explanations can additionally answer for which label, and against which alternative label, is a given input feature useful. We produce contrastive explanations via both high-level abstract concept attribution and low-level input token/span attribution for two NLP classification benchmarks. Our findings demonstrate the ability of label-contrastive explanations to provide fine-grained interpretability of model decisions.

The capacity of neural networks like the widely adopted transformer is known to be very high. Evidence is emerging that they learn successfully due to inductive bias in the training routine, typically a variant of gradient descent (GD). To better understand this bias, we study the tendency for transformer parameters to grow in magnitude (ℓ₂ norm) during training, and its implications for the emergent representations within self attention layers. Empirically, we document norm growth in the training of transformer language models, including T5 during its pretraining. As the parameters grow in magnitude, we prove that the network approximates a discretized network with saturated activation functions. Such “saturated” networks are known to have a reduced capacity compared to the full network family that can be described in terms of formal languages and automata. Our results suggest saturation is a new characterization of an inductive bias implicit in GD of particular interest for NLP. We leverage the emergent discrete structure in a saturated transformer to analyze the role of different attention heads, finding that some focus locally on a small number of positions, while other heads compute global averages, allowing counting. We believe understanding the interplay between these two capabilities may shed further light on the structure of computation within large transformers.

The Winograd Schema (WS) has been proposed as a test for measuring commonsense capabilities of models. Recently, pre-trained language model-based approaches have boosted performance on some WS benchmarks but the source of improvement is still not clear. This paper suggests that the apparent progress on WS may not necessarily reflect progress in commonsense reasoning. To support this claim, we first show that the current evaluation method of WS is sub-optimal and propose a modification that uses twin sentences for evaluation. We also propose two new baselines that indicate the existence of artifacts in WS benchmarks. We then develop a method for evaluating WS-like sentences in a zero-shot setting to account for the commonsense reasoning abilities acquired during the pretraining and observe that popular language models perform randomly in this setting when using our more strict evaluation. We conclude that the observed progress is mostly due to the use of supervision in training WS models, which is not likely to successfully support all the required commonsense reasoning skills and knowledge.

We introduce a large set of Hebrew lexicons pertaining to psychological aspects. These lexicons are useful for various psychology applications such as detecting emotional state, well being, relationship quality in conversation, identifying topics (e.g., family, work) and many more. We discuss the challenges in creating and validating lexicons in a new language, and highlight our methodological considerations in the data-driven lexicon construction process. Most of the lexicons are publicly available, which will facilitate further research on Hebrew clinical psychology text analysis. The lexicons were developed through data driven means, and verified by domain experts, clinical psychologists and psychology students, in a process of reconciliation with three judges. Development and verification relied on a dataset of a total of 872 psychotherapy session transcripts. We describe the construction process of each collection, the final resource and initial results of research studies employing this resource.

Recent advancements in self-attention neural network architectures have raised the bar for open-ended text generation. Yet, while current methods are capable of producing a coherent text which is several hundred words long, attaining control over the content that is being generated—as well as evaluating it—are still open questions. We propose a controlled generation task which is based on expanding a sequence of facts, expressed in natural language, into a longer narrative. We introduce human-based evaluation metrics for this task, as well as a method for deriving a large training dataset. We evaluate three methods on this task, based on fine-tuning pre-trained models. We show that while auto-regressive, unidirectional Language Models such as GPT2 produce better fluency, they struggle to adhere to the requested facts. We propose a plan-and-cloze model (using fine-tuned XLNet) which produces competitive fluency while adhering to the requested content.

Large pre-trained language models reach state-of-the-art results on many different NLP tasks when fine-tuned individually; They also come with a significant memory and computational requirements, calling for methods to reduce model sizes (green AI). We propose a two-stage model-compression method to reduce a model’s inference time cost. We first decompose the matrices in the model into smaller matrices and then perform feature distillation on the internal representation to recover from the decomposition. This approach has the benefit of reducing the number of parameters while preserving much of the information within the model. We experimented on BERT-base model with the GLUE benchmark dataset and show that we can reduce the number of parameters by a factor of 0.4x, and increase inference speed by a factor of 1.45x, while maintaining a minimal loss in metric performance.

We develop a formal hierarchy of the expressive capacity of RNN architectures. The hierarchy is based on two formal properties: space complexity, which measures the RNN’s memory, and rational recurrence, defined as whether the recurrent update can be described by a weighted finite-state machine. We place several RNN variants within this hierarchy. For example, we prove the LSTM is not rational, which formally separates it from the related QRNN (Bradbury et al., 2016). We also show how these models’ expressive capacity is expanded by stacking multiple layers or composing them with different pooling functions. Our results build on the theory of “saturated” RNNs (Merrill, 2019). While formally extending these findings to unsaturated RNNs is left to future work, we hypothesize that the practical learnable capacity of unsaturated RNNs obeys a similar hierarchy. We provide empirical results to support this conjecture. Experimental findings from training unsaturated networks on formal languages support this conjecture.

The problem of comparing two bodies of text and searching for words that differ in their usage between them arises often in digital humanities and computational social science. This is commonly approached by training word embeddings on each corpus, aligning the vector spaces, and looking for words whose cosine distance in the aligned space is large. However, these methods often require extensive filtering of the vocabulary to perform well, and - as we show in this work - result in unstable, and hence less reliable, results. We propose an alternative approach that does not use vector space alignment, and instead considers the neighbors of each word. The method is simple, interpretable and stable. We demonstrate its effectiveness in 9 different setups, considering different corpus splitting criteria (age, gender and profession of tweet authors, time of tweet) and different languages (English, French and Hebrew).

With the growing popularity of deep-learning based NLP models, comes a need for interpretable systems. But what is interpretability, and what constitutes a high-quality interpretation? In this opinion piece we reflect on the current state of interpretability evaluation research. We call for more clearly differentiating between different desired criteria an interpretation should satisfy, and focus on the faithfulness criteria. We survey the literature with respect to faithfulness evaluation, and arrange the current approaches around three assumptions, providing an explicit form to how faithfulness is “defined” by the community. We provide concrete guidelines on how evaluation of interpretation methods should and should not be conducted. Finally, we claim that the current binary definition for faithfulness sets a potentially unrealistic bar for being considered faithful. We call for discarding the binary notion of faithfulness in favor of a more graded one, which we believe will be of greater practical utility.

We tackle the task of Term Set Expansion (TSE): given a small seed set of example terms from a semantic class, finding more members of that class. The task is of great practical utility, and also of theoretical utility as it requires generalization from few examples. Previous approaches to the TSE task can be characterized as either distributional or pattern-based. We harness the power of neural masked language models (MLM) and propose a novel TSE algorithm, which combines the pattern-based and distributional approaches. Due to the small size of the seed set, fine-tuning methods are not effective, calling for more creative use of the MLM. The gist of the idea is to use the MLM to first mine for informative patterns with respect to the seed set, and then to obtain more members of the seed class by generalizing these patterns. Our method outperforms state-of-the-art TSE algorithms. Implementation is available at: https://github.com/guykush/TermSetExpansion-MPB/

The ability to control for the kinds of information encoded in neural representation has a variety of use cases, especially in light of the challenge of interpreting these models. We present Iterative Null-space Projection (INLP), a novel method for removing information from neural representations. Our method is based on repeated training of linear classifiers that predict a certain property we aim to remove, followed by projection of the representations on their null-space. By doing so, the classifiers become oblivious to that target property, making it hard to linearly separate the data according to it. While applicable for multiple uses, we evaluate our method on bias and fairness use-cases, and show that our method is able to mitigate bias in word embeddings, as well as to increase fairness in a setting of multi-class classification.

The notion of “in-domain data” in NLP is often over-simplistic and vague, as textual data varies in many nuanced linguistic aspects such as topic, style or level of formality. In addition, domain labels are many times unavailable, making it challenging to build domain-specific systems. We show that massive pre-trained language models implicitly learn sentence representations that cluster by domains without supervision – suggesting a simple data-driven definition of domains in textual data. We harness this property and propose domain data selection methods based on such models, which require only a small set of in-domain monolingual data. We evaluate our data selection methods for neural machine translation across five diverse domains, where they outperform an established approach as measured by both BLEU and precision and recall with respect to an oracle selection.

We present a system that allows a user to search a large linguistically annotated corpus using syntactic patterns over dependency graphs. In contrast to previous attempts to this effect, we introduce a light-weight query language that does not require the user to know the details of the underlying syntactic representations, and instead to query the corpus by providing an example sentence coupled with simple markup. Search is performed at an interactive speed due to efficient linguistic graph-indexing and retrieval engine. This allows for rapid exploration, development and refinement of syntax-based queries. We demonstrate the system using queries over two corpora: the English wikipedia, and a collection of English pubmed abstracts. A demo of the wikipedia system is available at https://allenai.github.io/spike/ .

Syntactic dependencies can be predicted with high accuracy, and are useful for both machine-learned and pattern-based information extraction tasks. However, their utility can be improved. These syntactic dependencies are designed to accurately reflect syntactic relations, and they do not make semantic relations explicit. Therefore, these representations lack many explicit connections between content words, that would be useful for downstream applications. Proposals like English Enhanced UD improve the situation by extending universal dependency trees with additional explicit arcs. However, they are not available to Python users, and are also limited in coverage. We introduce a broad-coverage, data-driven and linguistically sound set of transformations, that makes event-structure and many lexical relations explicit. We present pyBART, an easy-to-use open-source Python library for converting English UD trees either to Enhanced UD graphs or to our representation. The library can work as a standalone package or be integrated within a spaCy NLP pipeline. When evaluated in a pattern-based relation extraction scenario, our representation results in higher extraction scores than Enhanced UD, while requiring fewer patterns.

We present a system for automatic diacritization of Hebrew Text. The system combines modern neural models with carefully curated declarative linguistic knowledge and comprehensive manually constructed tables and dictionaries. Besides providing state of the art diacritization accuracy, the system also supports an interface for manual editing and correction of the automatic output, and has several features which make it particularly useful for preparation of scientific editions of historical Hebrew texts. The system supports Modern Hebrew, Rabbinic Hebrew and Poetic Hebrew. The system is freely accessible for all use at http://nakdanpro.dicta.org.il

Crowdsourcing has eased and scaled up the collection of linguistic annotation in recent years. In this work, we follow known methodologies of collecting labeled data for the complement coercion phenomenon. These are constructions with an implied action — e.g., “I started a new book I bought last week”, where the implied action is reading. We aim to collect annotated data for this phenomenon by reducing it to either of two known tasks: Explicit Completion and Natural Language Inference. However, in both cases, crowdsourcing resulted in low agreement scores, even though we followed the same methodologies as in previous work. Why does the same process fail to yield high agreement scores? We specify our modeling schemes, highlight the differences with previous work and provide some insights about the task and possible explanations for the failure. We conclude that specific phenomena require tailored solutions, not only in specialized algorithms, but also in data collection methods.

We present a system that allows life-science researchers to search a linguistically annotated corpus of scientific texts using patterns over dependency graphs, as well as using patterns over token sequences and a powerful variant of boolean keyword queries. In contrast to previous attempts to dependency-based search, we introduce a light-weight query language that does not require the user to know the details of the underlying linguistic representations, and instead to query the corpus by providing an example sentence coupled with simple markup. Search is performed at an interactive speed due to efficient linguistic graph-indexing and retrieval engine. This allows for rapid exploration, development and refinement of user queries. We demonstrate the system using example workflows over two corpora: the PubMed corpus including 14,446,243 PubMed abstracts and the CORD-19 dataset, a collection of over 45,000 research papers focused on COVID-19 research. The system is publicly available at https://allenai.github.io/spike

Understanding natural language questions entails the ability to break down a question into the requisite steps for computing its answer. In this work, we introduce a Question Decomposition Meaning Representation (QDMR) for questions. QDMR constitutes the ordered list of steps, expressed through natural language, that are necessary for answering a question. We develop a crowdsourcing pipeline, showing that quality QDMRs can be annotated at scale, and release the Break dataset, containing over 83K pairs of questions and their QDMRs. We demonstrate the utility of QDMR by showing that (a) it can be used to improve open-domain question answering on the HotpotQA dataset, (b) it can be deterministically converted to a pseudo-SQL formal language, which can alleviate annotation in semantic parsing applications. Last, we use Break to train a sequence-to-sequence model with copying that parses questions into QDMR structures, and show that it substantially outperforms several natural baselines.

Recent success of pre-trained language models (LMs) has spurred widespread interest in the language capabilities that they possess. However, efforts to understand whether LM representations are useful for symbolic reasoning tasks have been limited and scattered. In this work, we propose eight reasoning tasks, which conceptually require operations such as comparison, conjunction, and composition. A fundamental challenge is to understand whether the performance of a LM on a task should be attributed to the pre-trained representations or to the process of fine-tuning on the task data. To address this, we propose an evaluation protocol that includes both zero-shot evaluation (no fine-tuning), as well as comparing the learning curve of a fine-tuned LM to the learning curve of multiple controls, which paints a rich picture of the LM capabilities. Our main findings are that: (a) different LMs exhibit qualitatively different reasoning abilities, e.g., RoBERTa succeeds in reasoning tasks where BERT fails completely; (b) LMs do not reason in an abstract manner and are context-dependent, e.g., while RoBERTa can compare ages, it can do so only when the ages are in the typical range of human ages; (c) On half of our reasoning tasks all models fail completely. Our findings and infrastructure can help future work on designing new datasets, models, and objective functions for pre-training.

Recent works have demonstrated that multilingual BERT (mBERT) learns rich cross-lingual representations, that allow for transfer across languages. We study the word-level translation information embedded in mBERT and present two simple methods that expose remarkable translation capabilities with no fine-tuning. The results suggest that most of this information is encoded in a non-linear way, while some of it can also be recovered with purely linear tools. As part of our analysis, we test the hypothesis that mBERT learns representations which contain both a language-encoding component and an abstract, cross-lingual component, and explicitly identify an empirical language-identity subspace within mBERT representations.

Contextualized word representations, such as ELMo and BERT, were shown to perform well on various semantic and syntactic task. In this work, we tackle the task of unsupervised disentanglement between semantics and structure in neural language representations: we aim to learn a transformation of the contextualized vectors, that discards the lexical semantics, but keeps the structural information. To this end, we automatically generate groups of sentences which are structurally similar but semantically different, and use metric-learning approach to learn a transformation that emphasizes the structural component that is encoded in the vectors. We demonstrate that our transformation clusters vectors in space by structural properties, rather than by lexical semantics. Finally, we demonstrate the utility of our distilled representations by showing that they outperform the original contextualized representations in a few-shot parsing setting.

The process of collecting and annotating training data may introduce distribution artifacts which may limit the ability of models to learn correct generalization behavior. We identify failure modes of SOTA relation extraction (RE) models trained on TACRED, which we attribute to limitations in the data annotation process. We collect and annotate a challenge-set we call Challenging RE (CRE), based on naturally occurring corpus examples, to benchmark this behavior. Our experiments with four state-of-the-art RE models show that they have indeed adopted shallow heuristics that do not generalize to the challenge-set data. Further, we find that alternative question answering modeling performs significantly better than the SOTA models on the challenge-set, despite worse overall TACRED performance. By adding some of the challenge data as training examples, the performance of the model improves. Finally, we provide concrete suggestion on how to improve RE data collection to alleviate this behavior.

We provide the first computational treatment of fused-heads constructions (FHs), focusing on the numeric fused-heads (NFHs). FHs constructions are noun phrases in which the head noun is missing and is said to be “fused” with its dependent modifier. This missing information is implicit and is important for sentence understanding. The missing references are easily filled in by humans but pose a challenge for computational models. We formulate the handling of FHs as a two stages process: Identification of the FH construction and resolution of the missing head. We explore the NFH phenomena in large corpora of English text and create (1) a data set and a highly accurate method for NFH identification; (2) a 10k examples (1 M tokens) crowd-sourced data set of NFH resolution; and (3) a neural baseline for the NFH resolution task. We release our code and data set, to foster further research into this challenging problem.

Many natural languages assign grammatical gender also to inanimate nouns in the language. In such languages, words that relate to the gender-marked nouns are inflected to agree with the noun’s gender. We show that this affects the word representations of inanimate nouns, resulting in nouns with the same gender being closer to each other than nouns with different gender. While “embedding debiasing” methods fail to remove the effect, we demonstrate that a careful application of methods that neutralize grammatical gender signals from the words’ context when training word embeddings is effective in removing it. Fixing the grammatical gender bias yields a positive effect on the quality of the resulting word embeddings, both in monolingual and cross-lingual settings. We note that successfully removing gender signals, while achievable, is not trivial to do and that a language-specific morphological analyzer, together with careful usage of it, are essential for achieving good results.

The problem of learning to translate between two vector spaces given a set of aligned points arises in several application areas of NLP. Current solutions assume that the lexicon which defines the alignment pairs is noise-free. We consider the case where the set of aligned points is allowed to contain an amount of noise, in the form of incorrect lexicon pairs and show that this arises in practice by analyzing the edited dictionaries after the cleaning process. We demonstrate that such noise substantially degrades the accuracy of the learned translation when using current methods. We propose a model that accounts for noisy pairs. This is achieved by introducing a generative model with a compatible iterative EM algorithm. The algorithm jointly learns the noise level in the lexicon, finds the set of noisy pairs, and learns the mapping between the spaces. We demonstrate the effectiveness of our proposed algorithm on two alignment problems: bilingual word embedding translation, and mapping between diachronic embedding spaces for recovering the semantic shifts of words across time periods.

Word embeddings are widely used in NLP for a vast range of tasks. It was shown that word embeddings derived from text corpora reflect gender biases in society. This phenomenon is pervasive and consistent across different word embedding models, causing serious concern. Several recent works tackle this problem, and propose methods for significantly reducing this gender bias in word embeddings, demonstrating convincing results. However, we argue that this removal is superficial. While the bias is indeed substantially reduced according to the provided bias definition, the actual effect is mostly hiding the bias, not removing it. The gender bias information is still reflected in the distances between “gender-neutralized” words in the debiased embeddings, and can be recovered from them. We present a series of experiments to support this claim, for two debiasing methods. We conclude that existing bias removal techniques are insufficient, and should not be trusted for providing gender-neutral modeling.

Data-to-text generation can be conceptually divided into two parts: ordering and structuring the information (planning), and generating fluent language describing the information (realization). Modern neural generation systems conflate these two steps into a single end-to-end differentiable system. We propose to split the generation process into a symbolic text-planning stage that is faithful to the input, followed by a neural generation stage that focuses only on realization. For training a plan-to-text generator, we present a method for matching reference texts to their corresponding text plans. For inference time, we describe a method for selecting high-quality text plans for new inputs. We implement and evaluate our approach on the WebNLG benchmark. Our results demonstrate that decoupling text planning from neural realization indeed improves the system’s reliability and adequacy while maintaining fluent output. We observe improvements both in BLEU scores and in manual evaluations. Another benefit of our approach is the ability to output diverse realizations of the same input, paving the way to explicit control over the generated text structure.

How do typological properties such as word order and morphological case marking affect the ability of neural sequence models to acquire the syntax of a language? Cross-linguistic comparisons of RNNs’ syntactic performance (e.g., on subject-verb agreement prediction) are complicated by the fact that any two languages differ in multiple typological properties, as well as by differences in training corpus. We propose a paradigm that addresses these issues: we create synthetic versions of English, which differ from English in one or more typological parameters, and generate corpora for those languages based on a parsed English corpus. We report a series of experiments in which RNNs were trained to predict agreement features for verbs in each of those synthetic languages. Among other findings, (1) performance was higher in subject-verb-object order (as in English) than in subject-object-verb order (as in Japanese), suggesting that RNNs have a recency bias; (2) predicting agreement with both subject and object (polypersonal agreement) improves over predicting each separately, suggesting that underlying syntactic knowledge transfers across the two tasks; and (3) overt morphological case makes agreement prediction significantly easier, regardless of word order.

Deep learning systems thrive on abundance of labeled training data but such data is not always available, calling for alternative methods of supervision. One such method is expectation regularization (XR) (Mann and McCallum, 2007), where models are trained based on expected label proportions. We propose a novel application of the XR framework for transfer learning between related tasks, where knowing the labels of task A provides an estimation of the label proportion of task B. We then use a model trained for A to label a large corpus, and use this corpus with an XR loss to train a model for task B. To make the XR framework applicable to large-scale deep-learning setups, we propose a stochastic batched approximation procedure. We demonstrate the approach on the task of Aspect-based Sentiment classification, where we effectively use a sentence-level sentiment predictor to train accurate aspect-based predictor. The method improves upon fully supervised neural system trained on aspect-level data, and is also cumulative with LM-based pretraining, as we demonstrate by improving a BERT-based Aspect-based Sentiment model.

Crowdsourcing has been the prevalent paradigm for creating natural language understanding datasets in recent years. A common crowdsourcing practice is to recruit a small number of high-quality workers, and have them massively generate examples. Having only a few workers generate the majority of examples raises concerns about data diversity, especially when workers freely generate sentences. In this paper, we perform a series of experiments showing these concerns are evident in three recent NLP datasets. We show that model performance improves when training with annotator identifiers as features, and that models are able to recognize the most productive annotators. Moreover, we show that often models do not generalize well to examples from annotators that did not contribute to the training set. Our findings suggest that annotator bias should be monitored during dataset creation, and that test set annotators should be disjoint from training set annotators.

We focus on the problem of language modeling for code-switched language, in the context of automatic speech recognition (ASR). Language modeling for code-switched language is challenging for (at least) three reasons: (1) lack of available large-scale code-switched data for training; (2) lack of a replicable evaluation setup that is ASR directed yet isolates language modeling performance from the other intricacies of the ASR system; and (3) the reliance on generative modeling. We tackle these three issues: we propose an ASR-motivated evaluation setup which is decoupled from an ASR system and the choice of vocabulary, and provide an evaluation dataset for English-Spanish code-switching. This setup lends itself to a discriminative training approach, which we demonstrate to work better than generative language modeling. Finally, we explore a variety of training protocols and verify the effectiveness of training with large amounts of monolingual data followed by fine-tuning with small amounts of code-switched data, for both the generative and discriminative cases.

Word embeddings are widely used in NLP for a vast range of tasks. It was shown that word embeddings derived from text corpora reflect gender biases in society, causing serious concern. Several recent works tackle this problem, and propose methods for significantly reducing this gender bias in word embeddings, demonstrating convincing results. However, we argue that this removal is superficial. While the bias is indeed substantially reduced according to the provided bias definition, the actual effect is mostly hiding the bias, not removing it. The gender bias information is still reflected in the distances between “gender-neutralized” words in the debiased embeddings, and can be recovered from them. We present a series of experiments to support this claim, for two debiasing methods. We conclude that existing bias removal techniques are insufficient, and should not be trusted for providing gender-neutral modeling.

Many natural languages assign grammatical gender also to inanimate nouns in the language. In such languages, words that relate to the gender-marked nouns are inflected to agree with the noun’s gender. We show that this affects the word representations of inanimate nouns, resulting in nouns with the same gender being closer to each other than nouns with different gender. While “embedding debiasing” methods fail to remove the effect, we demonstrate that a careful application of methods that neutralize grammatical gender signals from the words’ context when training word embeddings is effective in removing it. Fixing the grammatical gender bias results in a positive effect on the quality of the resulting word embeddings, both in monolingual and cross lingual settings. We note that successfully removing gender signals, while achievable, is not trivial to do and that a language-specific morphological analyzer, together with careful usage of it, are essential for achieving good results.

When translating from a language that does not morphologically mark information such as gender and number into a language that does, translation systems must “guess” this missing information, often leading to incorrect translations in the given context. We propose a black-box approach for injecting the missing information to a pre-trained neural machine translation system, allowing to control the morphological variations in the generated translations without changing the underlying model or training data. We evaluate our method on an English to Hebrew translation task, and show that it is effective in injecting the gender and number information and that supplying the correct information improves the translation accuracy in up to 2.3 BLEU on a female-speaker test set for a state-of-the-art online black-box system. Finally, we perform a fine-grained syntactic analysis of the generated translations that shows the effectiveness of our method.

We follow the step-by-step approach to neural data-to-text generation proposed by Moryossef et al (2019), in which the generation process is divided into a text planning stage followed by a plan realization stage. We suggest four extensions to that framework: (1) we introduce a trainable neural planning component that can generate effective plans several orders of magnitude faster than the original planner; (2) we incorporate typing hints that improve the model’s ability to deal with unseen relations and entities; (3) we introduce a verification-by-reranking stage that substantially improves the faithfulness of the resulting texts; (4) we incorporate a simple but effective referring expression generation module. These extensions result in a generation process that is faster, more fluent, and more accurate.

We create a new NLI test set that shows the deficiency of state-of-the-art models in inferences that require lexical and world knowledge. The new examples are simpler than the SNLI test set, containing sentences that differ by at most one word from sentences in the training set. Yet, the performance on the new test set is substantially worse across systems trained on SNLI, demonstrating that these systems are limited in their generalization ability, failing to capture many simple inferences.

Splitting and rephrasing a complex sentence into several shorter sentences that convey the same meaning is a challenging problem in NLP. We show that while vanilla seq2seq models can reach high scores on the proposed benchmark (Narayan et al., 2017), they suffer from memorization of the training set which contains more than 89% of the unique simple sentences from the validation and test sets. To aid this, we present a new train-development-test data split and neural models augmented with a copy-mechanism, outperforming the best reported baseline by 8.68 BLEU and fostering further progress on the task.

While Recurrent Neural Networks (RNNs) are famously known to be Turing complete, this relies on infinite precision in the states and unbounded computation time. We consider the case of RNNs with finite precision whose computation time is linear in the input length. Under these limitations, we show that different RNN variants have different computational power. In particular, we show that the LSTM and the Elman-RNN with ReLU activation are strictly stronger than the RNN with a squashing activation and the GRU. This is achieved because LSTMs and ReLU-RNNs can easily implement counting behavior. We show empirically that the LSTM does indeed learn to effectively use the counting mechanism.

We present SetExpander, a corpus-based system for expanding a seed set of terms into a more complete set of terms that belong to the same semantic class. SetExpander implements an iterative end-to end workflow for term set expansion. It enables users to easily select a seed set of terms, expand it, view the expanded set, validate it, re-expand the validated set and store it, thus simplifying the extraction of domain-specific fine-grained semantic classes. SetExpander has been used for solving real-life use cases including integration in an automated recruitment system and an issues and defects resolution system. A video demo of SetExpander is available at https://drive.google.com/open?id=1e545bB87Autsch36DjnJHmq3HWfSd1Rv .

We present an analysis into the inner workings of Convolutional Neural Networks (CNNs) for processing text. CNNs used for computer vision can be interpreted by projecting filters into image space, but for discrete sequence inputs CNNs remain a mystery. We aim to understand the method by which the networks process and classify text. We examine common hypotheses to this problem: that filters, accompanied by global max-pooling, serve as ngram detectors. We show that filters may capture several different semantic classes of ngrams by using different activation patterns, and that global max-pooling induces behavior which separates important ngrams from the rest. Finally, we show practical use cases derived from our findings in the form of model interpretability (explaining a trained model by deriving a concrete identity for each filter, bridging the gap between visualization tools in vision tasks and NLP) and prediction interpretability (explaining predictions).

Sequential neural networks models are powerful tools in a variety of Natural Language Processing (NLP) tasks. The sequential nature of these models raises the questions: to what extent can these models implicitly learn hierarchical structures typical to human language, and what kind of grammatical phenomena can they acquire? We focus on the task of agreement prediction in Basque, as a case study for a task that requires implicit understanding of sentence structure and the acquisition of a complex but consistent morphological system. Analyzing experimental results from two syntactic prediction tasks – verb number prediction and suffix recovery – we find that sequential models perform worse on agreement prediction in Basque than one might expect on the basis of a previous agreement prediction work in English. Tentative findings based on diagnostic classifiers suggest the network makes use of local heuristics as a proxy for the hierarchical structure of the sentence. We propose the Basque agreement prediction task as challenging benchmark for models that attempt to learn regularities in human language.

Recent advances in Representation Learning and Adversarial Training seem to succeed in removing unwanted features from the learned representation. We show that demographic information of authors is encoded in—and can be recovered from—the intermediate representations learned by text-based neural classifiers. The implication is that decisions of classifiers trained on textual data are not agnostic to—and likely condition on—demographic attributes. When attempting to remove such demographic information using adversarial training, we find that while the adversarial component achieves chance-level development-set accuracy during training, a post-hoc classifier, trained on the encoded sentences from the first part, still manages to reach substantially higher classification accuracies on the same data. This behavior is consistent across several tasks, demographic properties and datasets. We explore several techniques to improve the effectiveness of the adversarial component. Our main conclusion is a cautionary one: do not rely on the adversarial training to achieve invariant representation to sensitive features.

An established method for Word Sense Induction (WSI) uses a language model to predict probable substitutes for target words, and induces senses by clustering these resulting substitute vectors. We replace the ngram-based language model (LM) with a recurrent one. Beyond being more accurate, the use of the recurrent LM allows us to effectively query it in a creative way, using what we call dynamic symmetric patterns. The combination of the RNN-LM and the dynamic symmetric patterns results in strong substitute vectors for WSI, allowing to surpass the current state-of-the-art on the SemEval 2013 WSI shared task by a large margin.

Recent work has explored the syntactic abilities of RNNs using the subject-verb agreement task, which diagnoses sensitivity to sentence structure. RNNs performed this task well in common cases, but faltered in complex sentences (Linzen et al., 2016). We test whether these errors are due to inherent limitations of the architecture or to the relatively indirect supervision provided by most agreement dependencies in a corpus. We trained a single RNN to perform both the agreement task and an additional task, either CCG supertagging or language modeling. Multi-task training led to significantly lower error rates, in particular on complex sentences, suggesting that RNNs have the ability to evolve more sophisticated syntactic representations than shown before. We also show that easily available agreement training data can improve performance on other syntactic tasks, in particular when only a limited amount of training data is available for those tasks. The multi-task paradigm can also be leveraged to inject grammatical knowledge into language models.

We present the Uppsala submission to the CoNLL 2017 shared task on parsing from raw text to universal dependencies. Our system is a simple pipeline consisting of two components. The first performs joint word and sentence segmentation on raw text; the second predicts dependency trees from raw words. The parser bypasses the need for part-of-speech tagging, but uses word embeddings based on universal tag distributions. We achieved a macro-averaged LAS F1 of 65.11 in the official test run, which improved to 70.49 after bug fixes. We obtained the 2nd best result for sentence segmentation with a score of 89.03.

We introduce a greedy transition-based parser that learns to represent parser states using recurrent neural networks. Our primary innovation that enables us to do this efficiently is a new control structure for sequential neural networks—the stack long short-term memory unit (LSTM). Like the conventional stack data structures used in transition-based parsers, elements can be pushed to or popped from the top of the stack in constant time, but, in addition, an LSTM maintains a continuous space embedding of the stack contents. Our model captures three facets of the parser’s state: (i) unbounded look-ahead into the buffer of incoming words, (ii) the complete history of transition actions taken by the parser, and (iii) the complete contents of the stack of partially built tree fragments, including their internal structures. In addition, we compare two different word representations: (i) standard word vectors based on look-up tables and (ii) character-based models of words. Although standard word embedding models work well in all languages, the character-based models improve the handling of out-of-vocabulary words, particularly in morphologically rich languages. Finally, we discuss the use of dynamic oracles in training the parser. During training, dynamic oracles alternate between sampling parser states from the training data and from the model as it is being learned, making the model more robust to the kinds of errors that will be made at test time. Training our model with dynamic oracles yields a linear-time greedy parser with very competitive performance.

While cross-lingual word embeddings have been studied extensively in recent years, the qualitative differences between the different algorithms remain vague. We observe that whether or not an algorithm uses a particular feature set (sentence IDs) accounts for a significant performance gap among these algorithms. This feature set is also used by traditional alignment algorithms, such as IBM Model-1, which demonstrate similar performance to state-of-the-art embedding algorithms on a variety of benchmarks. Overall, we observe that different algorithmic approaches for utilizing the sentence ID feature space result in similar performance. This paper draws both empirical and theoretical parallels between the embedding and alignment literature, and suggests that adding additional sources of information, which go beyond the traditional signal of bilingual sentence-aligned corpora, may substantially improve cross-lingual word embeddings, and that future baselines should at least take such features into account.

While dependency parsers reach very high overall accuracy, some dependency relations are much harder than others. In particular, dependency parsers perform poorly in coordination construction (i.e., correctly attaching the conj relation). We extend a state-of-the-art dependency parser with conjunction-specific features, focusing on the similarity between the conjuncts head words. Training the extended parser yields an improvement in conj attachment as well as in overall dependency parsing accuracy on the Stanford dependency conversion of the Penn TreeBank.

We explore the ability of word embeddings to capture both semantic and morphological similarity, as affected by the different types of linguistic properties (surface form, lemma, morphological tag) used to compose the representation of each word. We train several models, where each uses a different subset of these properties to compose its representations. By evaluating the models on semantic and morphological measures, we reveal some useful insights on the relationship between semantics and morphology.

Most work on neural natural language generation (NNLG) focus on controlling the content of the generated text. We experiment with controling several stylistic aspects of the generated text, in addition to its content. The method is based on conditioned RNN language model, where the desired content as well as the stylistic parameters serve as conditioning contexts. We demonstrate the approach on the movie reviews domain and show that it is successful in generating coherent sentences corresponding to the required linguistic style and content.

We present a neural model for morphological inflection generation which employs a hard attention mechanism, inspired by the nearly-monotonic alignment commonly found between the characters in a word and the characters in its inflection. We evaluate the model on three previously studied morphological inflection generation datasets and show that it provides state of the art results in various setups compared to previous neural and non-neural approaches. Finally we present an analysis of the continuous representations learned by both the hard and soft (Bahdanau, 2014) attention models for the task, shedding some light on the features such models extract.

We present a simple method to incorporate syntactic information about the target language in a neural machine translation system by translating into linearized, lexicalized constituency trees. An experiment on the WMT16 German-English news translation task resulted in an improved BLEU score when compared to a syntax-agnostic NMT baseline trained on the same dataset. An analysis of the translations from the syntax-aware system shows that it performs more reordering during translation in comparison to the baseline. A small-scale human evaluation also showed an advantage to the syntax-aware system.

This tutorial aims to bring NLP researchers up to speed with the current techniques in deep learning and neural networks, and show them how they can turn their ideas into practical implementations. We will start with simple classification models (logistic regression and multilayer perceptrons) and cover more advanced patterns that come up in NLP such as recurrent networks for sequence tagging and prediction problems, structured networks (e.g., compositional architectures based on syntax trees), structured output spaces (sequences and trees), attention for sequence-to-sequence transduction, and feature induction for complex algorithm states. A particular emphasis will be on learning to represent complex objects as recursive compositions of simpler objects. This representation will reflect characterize standard objects in NLP, such as the composition of characters and morphemes into words, and words into sentences and documents. In addition, new opportunities such as learning to embed "algorithm states" such as those used in transition-based parsing and other sequential structured prediction models (for which effective features may be difficult to engineer by hand) will be covered.Everything in the tutorial will be grounded in code — we will show how to program seemingly complex neural-net models using toolkits based on the computation-graph formalism. Computation graphs decompose complex computations into a DAG, with nodes representing inputs, target outputs, parameters, or (sub)differentiable functions (e.g., "tanh", "matrix multiply", and "softmax"), and edges represent data dependencies. These graphs can be run "forward" to make predictions and compute errors (e.g., log loss, squared error) and then "backward" to compute derivatives with respect to model parameters. In particular we'll cover the Python bindings of the CNN library. CNN has been designed from the ground up for NLP applications, dynamically structured NNs, rapid prototyping, and a transparent data and execution model.

We present a simple and effective scheme for dependency parsing which is based on bidirectional-LSTMs (BiLSTMs). Each sentence token is associated with a BiLSTM vector representing the token in its sentential context, and feature vectors are constructed by concatenating a few BiLSTM vectors. The BiLSTM is trained jointly with the parser objective, resulting in very effective feature extractors for parsing. We demonstrate the effectiveness of the approach by applying it to a greedy transition-based parser as well as to a globally optimized graph-based parser. The resulting parsers have very simple architectures, and match or surpass the state-of-the-art accuracies on English and Chinese.

We suggest a compositional vector representation of parse trees that relies on a recursive combination of recurrent-neural network encoders. To demonstrate its effectiveness, we use the representation as the backbone of a greedy, bottom-up dependency parser, achieving very strong accuracies for English and Chinese, without relying on external word embeddings. The parser’s implementation is available for download at the first author’s webpage.

The success of long short-term memory (LSTM) neural networks in language processing is typically attributed to their ability to capture long-distance statistical regularities. Linguistic regularities are often sensitive to syntactic structure; can such dependencies be captured by LSTMs, which do not have explicit structural representations? We begin addressing this question using number agreement in English subject-verb dependencies. We probe the architecture’s grammatical competence both using training objectives with an explicit grammatical target (number prediction, grammaticality judgments) and using language models. In the strongly supervised settings, the LSTM achieved very high overall accuracy (less than 1% errors), but errors increased when sequential and structural information conflicted. The frequency of such errors rose sharply in the language-modeling setting. We conclude that LSTMs can capture a non-trivial amount of grammatical structure given targeted supervision, but stronger architectures may be required to further reduce errors; furthermore, the language modeling signal is insufficient for capturing syntax-sensitive dependencies, and should be supplemented with more direct supervision if such dependencies need to be captured.

Cross-linguistically consistent annotation is necessary for sound comparative evaluation and cross-lingual learning experiments. It is also useful for multilingual system development and comparative linguistic studies. Universal Dependencies is an open community effort to create cross-linguistically consistent treebank annotation for many languages within a dependency-based lexicalist framework. In this paper, we describe v1 of the universal guidelines, the underlying design principles, and the currently available treebanks for 33 languages.

Prepositions are very common and very ambiguous, and understanding their sense is critical for understanding the meaning of the sentence. Supervised corpora for the preposition-sense disambiguation task are small, suggesting a semi-supervised approach to the task. We show that signals from unannotated multilingual data can be used to improve supervised preposition-sense disambiguation. Our approach pre-trains an LSTM encoder for predicting the translation of a preposition, and then incorporates the pre-trained encoder as a component in a supervised classification system, and fine-tunes it for the task. The multilingual signals consistently improve results on two preposition-sense datasets.

Recent trends suggest that neural-network-inspired word embedding models outperform traditional count-based distributional models on word similarity and analogy detection tasks. We reveal that much of the performance gains of word embeddings are due to certain system design choices and hyperparameter optimizations, rather than the embedding algorithms themselves. Furthermore, we show that these modifications can be transferred to traditional distributional models, yielding similar gains. In contrast to prior reports, we observe mostly local or insignificant performance differences between the methods, with no global advantage to any single approach over the others.

We develop parsing oracles for two transition-based dependency parsers, including the arc-standard parser, solving a problem that was left open in (Goldberg and Nivre, 2013). We experimentally show that using these oracles during training yields superior parsing accuracies on many languages.

Greedy transition-based parsers are very fast but tend to suffer from error propagation. This problem is aggravated by the fact that they are normally trained using oracles that are deterministic and incomplete in the sense that they assume a unique canonical path through the transition system and are only valid as long as the parser does not stray from this path. In this paper, we give a general characterization of oracles that are nondeterministic and complete, present a method for deriving such oracles for transition systems that satisfy a property we call arc decomposition, and instantiate this method for three well-known transition systems from the literature. We say that these oracles are dynamic, because they allow us to dynamically explore alternative and nonoptimal paths during training — in contrast to oracles that statically assume a unique optimal path. Experimental evaluation on a wide range of data sets clearly shows that using dynamic oracles to train greedy parsers gives substantial improvements in accuracy. Moreover, this improvement comes at no cost in terms of efficiency, unlike other techniques like beam search.

We report on an effort to build a corpus of Modern Hebrew tagged with part-of-speech and morphology. We designed a tagset specific to Hebrew while focusing on four aspects: the tagset should be consistent with common linguistic knowledge; there should be maximal agreement among taggers as to the tags assigned to maintain consistency; the tagset should be useful for machine taggers and learning algorithms; and the tagset should be effective for applications relying on the tags as input features. In this paper, we illustrate these issues by explaining our decision to introduce a tag for beinoni forms in Hebrew. We explain how this tag is defined, and how it helped us improve manual tagging accuracy to a high-level, while improving automatic tagging and helping in the task of syntactic chunking.

Morphologically rich languages pose a challenge to the annotators of treebanks with respect to the status of orthographic (space-delimited) words in the syntactic parse trees. In such languages an orthographic word may carry various, distinct, sorts of information and the question arises whether we should represent such words as a sequence of their constituent morphemes (i.e., a Morpheme-Based annotation strategy) or whether we should preserve their special orthographic status within the trees (i.e., a Word-Based annotation strategy). In this paper we empirically address this challenge in the context of the development of Language Resources for Modern Hebrew. We compare and contrast the Morpheme-Based and Word-Based annotation strategies of pronominal clitics in Modern Hebrew and we show that the Word-Based strategy is more adequate for the purpose of training statistical parsers as it provides a better PP-attachment disambiguation capacity and a better alignment with initial surface forms. Our findings in turn raise new questions concerning the interaction of morphological and syntactic processing of which investigation is facilitated by the parallel treebank we made available.