We deploy the methods of controlled psycholinguistic experimentation to shed light on the extent to which the behavior of neural network language models reflects incremental representations of syntactic state. To do so, we examine model behavior on artificial sentences containing a variety of syntactically complex structures. We test four models: two publicly available LSTM sequence models of English (Jozefowicz et al., 2016; Gulordava et al., 2018) trained on large datasets; an RNNG (Dyer et al., 2016) trained on a small, parsed dataset; and an LSTM trained on the same small corpus as the RNNG. We find evidence that the LSTMs trained on large datasets represent syntactic state over large spans of text in a way that is comparable to the RNNG, while the LSTM trained on the small dataset does not or does so only weakly.
UC Irvine（加州大学欧文分校）；Harvard University（哈佛大学）；Primate Research Institute, Kyoto University（京都大学灵长类研究所）；MIT（麻省理工学院）
Recurrent neural networks (RNNs) are the state of the art in sequence modeling for natural language. However, it remains poorly understood what grammatical characteristics of natural language they implicitly learn and represent as a consequence of optimizing the language modeling objective. Here we deploy the methods of controlled psycholinguistic experimentation to shed light on to what extent RNN behavior reflects incremental syntactic state and grammatical dependency representations known to characterize human linguistic behavior. We broadly test two publicly available long short-term memor…
Centro Singular de Investigación en Tecnoloxías Intelixentes (CiTIUS)（智能技术单一研究中心（CiTIUS））；Universidade de Santiago de Compostela（圣地亚哥德孔波斯特拉大学）
This paper explores the ability of Transformer models to capture subject-verb and noun-adjective agreement dependencies in Galician. We conduct a series of word prediction experiments in which we manipulate dependency length together with the presence of an attractor noun that acts as a lure. First, we evaluate the overall performance of the existing monolingual and multilingual models for Galician. Secondly, to observe the effects of the training process, we compare the different degrees of achievement of two monolingual BERT models at different training points. We also release their checkpoi…
University of São Paulo（圣保罗大学）；Federal Technological University of Paraná (UTFPR)（巴拉那联邦理工大学）；University of Sheffield（谢菲尔德大学）
Psycholinguistic properties of words have been used in various approaches to Natural Language Processing tasks, such as text simplification and readability assessment. Most of these properties are subjective, involving costly and time-consuming surveys to be gathered. Recent approaches use the limited datasets of psycholinguistic properties to extend them automatically to large lexicons. However, some of the resources used by such approaches are not available to most languages. This study presents a method to infer psycholinguistic properties for Brazilian Portuguese (BP) using regressors buil…
Neural language models (LMs) perform well on tasks that require sensitivity to syntactic structure. Drawing on the syntactic priming paradigm from psycholinguistics, we propose a novel technique to analyze the representations that enable such success. By establishing a gradient similarity metric between structures, this technique allows us to reconstruct the organization of the LMs' syntactic representational space. We use this technique to demonstrate that LSTM LMs' representations of different types of sentences with relative clauses are organized hierarchically in a linguistically interpret…
This thesis presents a broad-coverage probabilistic top-down parser, and its application to the problem of language modeling for speech recognition. The parser builds fully connected derivations incrementally, in a single pass from left-to-right across the string. We argue that the parsing approach that we have adopted is well-motivated from a psycholinguistic perspective, as a model that captures probabilistic dependencies between lexical items, as part of the process of building connected syntactic structures. The basic parser and conditional probability models are presented, and empirical r…
It has been argued that humans rapidly adapt their lexical and syntactic expectations to match the statistics of the current linguistic context. We provide further support to this claim by showing that the addition of a simple adaptation mechanism to a neural language model improves our predictions of human reading times compared to a non-adaptive model. We analyze the performance of the model on controlled materials from psycholinguistic experiments and show that it adapts not only to lexical items but also to abstract syntactic structures.
Language models (LMs) trained on large quantities of text have been claimed to acquire abstract linguistic representations. Our work tests the robustness of these abstractions by focusing on the ability of LMs to learn interactions between different linguistic representations. In particular, we utilized stimuli from psycholinguistic studies showing that humans can condition reference (i.e. coreference resolution) and syntactic processing on the same discourse structure (implicit causality). We compared both transformer and long short-term memory LMs to find that, contrary to humans, implicit c…
École Normale Supérieure-PSL（巴黎高等师范学院（PSL大学））；CNRS（法国国家科学研究中心）；U. Sorbonne Nouvelle（新索邦大学）；University of Pisa（比萨大学）
Both humans and neural language models are able to perform subject-verb number agreement (SVA). In principle, semantics shouldn't interfere with this task, which only requires syntactic knowledge. In this work we test whether meaning interferes with this type of agreement in English in syntactic structures of various complexities. To do so, we generate both semantically well-formed and nonsensical items. We compare the performance of BERT-base to that of humans, obtained with a psycholinguistic online crowdsourcing experiment. We find that BERT and humans are both sensitive to our semantic man…
By positing a relationship between naturalistic reading times and information-theoretic surprisal, surprisal theory (Hale, 2001; Levy, 2008) provides a natural interface between language models and psycholinguistic models. This paper re-evaluates a claim due to Goodkind and Bicknell (2018) that a language model's ability to model reading times is a linear function of its perplexity. By extending Goodkind and Bicknell's analysis to modern neural architectures, we show that the proposed relation does not always hold for Long Short-Term Memory networks, Transformers, and pre-trained models. We in…
This paper reports on the comparison of the subject and object of verbs in their usage between phishing emails and legitimate emails. The purpose of this research is to explore whether the syntactic structures and subjects and objects of verbs can be distinguishable features for phishing detection. To achieve the objective, we have conducted two series of experiments: the syntactic similarity for sentences, and the subject and object of verb comparison. The results of the experiments indicated that both features can be used for some verbs, but more work has to be done for others.
Computer Science Department, Bar Ilan University（巴伊兰大学计算机科学系）；Allen Institute for Artificial Intelligence（艾伦人工智能研究所）
I assess the extent to which the recently introduced BERT model captures English syntactic phenomena, using (1) naturally-occurring subject-verb agreement stimuli; (2) "coloreless green ideas" subject-verb agreement stimuli, in which content words in natural sentences are randomly replaced with words sharing the same part-of-speech and inflection; and (3) manually crafted stimuli for subject-verb agreement and reflexive anaphora phenomena. The BERT model performs remarkably well on all cases.
SOKENDAI Graduate University for Advanced Studies（综合研究大学院大学）
The primary goal in this thesis is to identify better syntactic constraint or bias, that is language independent but also efficiently exploitable during sentence processing. We focus on a particular syntactic construction called center-embedding, which is well studied in psycholinguistics and noted to cause particular difficulty for comprehension. Since people use language as a tool for communication, one expects such complex constructions to be avoided for communication efficiency. From a computational perspective, center-embedding is closely relevant to a left-corner parsing algorithm, which…
Harvard University（哈佛大学）；Johns Hopkins University（约翰斯·霍普金斯大学）；Google Research（谷歌研究院）；New York University（纽约大学）；Technion – IIT（以色列理工学院）
Targeted syntactic evaluations have demonstrated the ability of language models to perform subject-verb agreement given difficult contexts. To elucidate the mechanisms by which the models accomplish this behavior, this study applies causal mediation analysis to pre-trained neural language models. We investigate the magnitude of models' preferences for grammatical inflections, as well as whether neurons process subject-verb agreement similarly across sentences with different syntactic structures. We uncover similarities and differences across architectures and model sizes -- notably, that large…
I review evidence for the claim that syntactic ambiguities are resolved on the basis of the meaning of the competing analyses, not their structure. I identify a collection of ambiguities that do not yet have a meaning-based account and propose one which is based on the interaction of discourse and grammatical function. I provide evidence for my proposal by examining statistical properties of the Penn Treebank of syntactically annotated text.
McGovern Medical School, University of Texas Health Science Center（德克萨斯大学健康科学中心麦戈文医学院）；Texas Institute for Restorative Neurotechnologies, University of Texas Health Science Center（德克萨斯大学健康科学中心德克萨斯恢复性神经技术研究所）；University College London（伦敦大学学院）；UCL Queen Square Institute of Neurology（伦敦大学学院皇后广场神经病学研究所）
Natural language syntax yields an unbounded array of hierarchically structured expressions. We claim that these are used in the service of active inference in accord with the free-energy principle (FEP). While conceptual advances alongside modelling and simulation work have attempted to connect speech segmentation and linguistic communication with the FEP, we extend this program to the underlying computations responsible for generating syntactic objects. We argue that recently proposed principles of economy in language design - such as "minimal search" criteria from theoretical syntax - adhere…
ALTA Institute（ALTA研究所）；University of Cambridge（剑桥大学）；Imperial College London（帝国理工学院）
Targeted studies testing knowledge of subject-verb agreement (SVA) indicate that pre-trained language models encode syntactic information. We assert that if models robustly encode subject-verb agreement, they should be able to identify when agreement is correct and when it is incorrect. To that end, we propose grammatical error detection as a diagnostic probe to evaluate token-level contextual representations for their knowledge of SVA. We evaluate contextual representations at each layer from five pre-trained English language models: BERT, XLNet, GPT-2, RoBERTa, and ELECTRA. We leverage publi…
École Normale Supérieure-PSL（巴黎高等师范学院-PSL大学）；CNRS（法国国家科学研究中心）；U. Sorbonne Nouvelle（新索邦大学）；University of Pisa（比萨大学）
Although transformer-based Neural Language Models demonstrate impressive performance on a variety of tasks, their generalization abilities are not well understood. They have been shown to perform strongly on subject-verb number agreement in a wide array of settings, suggesting that they learned to track syntactic dependencies during their training even without explicit supervision. In this paper, we examine the extent to which BERT is able to perform lexically-independent subject-verb number agreement (NA) on targeted syntactic templates. To do so, we disrupt the lexical patterns found in natu…
Harvard University（哈佛大学）；Massachusetts Institute of Technology（麻省理工学院）
Human reading behavior is tuned to the statistics of natural language: the time it takes human subjects to read a word can be predicted from estimates of the word's probability in context. However, it remains an open question what computational architecture best characterizes the expectations deployed in real time by humans that determine the behavioral signatures of reading. Here we test over two dozen models, independently manipulating computational architecture and training dataset size, on how well their next-word expectations predict human reading time behavior on naturalistic text corpor…
Université Paris Cité（巴黎西岱大学）；LLF（语言学形式实验室）；CNRS（法国国家科学研究中心）
The long-distance agreement, evidence for syntactic structure, is increasingly used to assess the syntactic generalization of Neural Language Models. Much work has shown that transformers are capable of high accuracy in varied agreement tasks, but the mechanisms by which the models accomplish this behavior are still not well understood. To better understand transformers' internal working, this work contrasts how they handle two superficially similar but theoretically distinct agreement phenomena: subject-verb and object-past participle agreement in French. Using probing and counterfactual anal…
Johns Hopkins University（约翰斯·霍普金斯大学）；New York University（纽约大学）
Structural probing work has found evidence for latent syntactic information in pre-trained language models. However, much of this analysis has focused on monolingual models, and analyses of multilingual models have employed correlational methods that are confounded by the choice of probing tasks. In this study, we causally probe multilingual language models (XGLM and multilingual BERT) as well as monolingual BERT-based models across various languages; we do this by performing counterfactual perturbations on neuron activations and observing the effect on models' subject-verb agreement probabili…
Recurrent Neural Networks (RNNs) trained on a language modeling task have been shown to acquire a number of non-local grammatical dependencies with some success. Here, we provide new evidence that RNN language models are sensitive to hierarchical syntactic structure by investigating the filler--gap dependency and constraints on it, known as syntactic islands. Previous work is inconclusive about whether RNNs learn to attenuate their expectations for gaps in island constructions in particular or in any sufficiently complex syntactic environment. This paper gives new evidence for the former by pr…
Signal Analysis and Interpretation Laboratory（信号分析与解释实验室）；Department of Electrical and Computer Engineering（电气与计算机工程系）；University of Southern California（南加州大学）
Most current language modeling techniques only exploit co-occurrence, semantic and syntactic information from the sequence of words. However, a range of information such as the state of the speaker and dynamics of the interaction might be useful. In this work we derive motivation from psycholinguistics and propose the addition of behavioral information into the context of language modeling. We propose the augmentation of language models with an additional module which analyzes the behavioral state of the current context. This behavioral information is used to gate the outputs of the language m…
Determining the correct form of a verb in context requires an understanding of the syntactic structure of the sentence. Recurrent neural networks have been shown to perform this task with an error rate comparable to humans, despite the fact that they are not designed with explicit syntactic representations. To examine the extent to which the syntactic representations of these networks are similar to those used by humans when processing sentences, we compare the detailed pattern of errors that RNNs and humans make on this task. Despite significant similarities (attraction errors, asymmetry betw…
Targeted syntactic evaluation of subject-verb number agreement in English (TSE) evaluates language models' syntactic knowledge using hand-crafted minimal pairs of sentences that differ only in the main verb's conjugation. The method evaluates whether language models rate each grammatical sentence as more likely than its ungrammatical counterpart. We identify two distinct goals for TSE. First, evaluating the systematicity of a language model's syntactic knowledge: given a sentence, can it conjugate arbitrary verbs correctly? Second, evaluating a model's likely behavior: given a sentence, does t…
Stanford University（斯坦福大学）；Saarland University（萨尔大学）；University of Toronto（多伦多大学）
Languages vary considerably in syntactic structure. About 40% of the world's languages have subject-verb-object order, and about 40% have subject-object-verb order. Extensive work has sought to explain this word order variation across languages. However, the existing approaches are not able to explain coherently the frequency distribution and evolution of word order in individual languages. We propose that variation in word order reflects different ways of balancing competing pressures of dependency locality and information locality, whereby languages favor placing elements together when they …
We present a dataset for evaluating the grammaticality of the predictions of a language model. We automatically construct a large number of minimally different pairs of English sentences, each consisting of a grammatical and an ungrammatical sentence. The sentence pairs represent different variations of structure-sensitive phenomena: subject-verb agreement, reflexive anaphora and negative polarity items. We expect a language model to assign a higher probability to the grammatical sentence than the ungrammatical one. In an experiment using this data set, an LSTM language model performed poorly …
A range of studies have concluded that neural word prediction models can distinguish grammatical from ungrammatical sentences with high accuracy. However, these studies are based primarily on monolingual evidence from English. To investigate how these models' ability to learn syntax varies by language, we introduce CLAMS (Cross-Linguistic Assessment of Models on Syntax), a syntactic evaluation suite for monolingual and multilingual models. CLAMS includes subject-verb agreement challenge sets for English, French, German, Hebrew and Russian, generated from grammars we develop. We use CLAMS to ev…
PSL Research University（巴黎文理研究大学）；Bar Ilan University（巴伊兰大学）；CNRS（法国国家科学研究中心）；EHESS（法国社会科学高等研究院）；Johns Hopkins University（约翰斯·霍普金斯大学）
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 low…
Artificial Intelligence Research Center（人工智能研究中心）；AIST（产业技术综合研究所）
We explore the utilities of explicit negative examples in training neural language models. Negative examples here are incorrect words in a sentence, such as "barks" in "*The dogs barks". Neural language models are commonly trained only on positive examples, a set of sentences in the training data, but recent studies suggest that the models trained in this way are not capable of robustly handling complex syntactic constructions, such as long-distance agreement. In this paper, using English data, we first demonstrate that appropriately using negative examples about particular constructions (e.g.…
While there has been significant progress towards developing NLU resources for Indic languages, syntactic evaluation has been relatively less explored. Unlike English, Indic languages have rich morphosyntax, grammatical genders, free linear word-order, and highly inflectional morphology. In this paper, we introduce Vyākarana: a benchmark of Colorless Green sentences in Indic languages for syntactic evaluation of multilingual language models. The benchmark comprises four syntax-related tasks: PoS Tagging, Syntax Tree-depth Prediction, Grammatical Case Marking, and Subject-Verb Agreement. We use…
Nara Institute of Science and Technology（奈良先端科学技术大学院大学）；Graduate School of Arts and Sciences, The University of Tokyo（东京大学综合文化研究科）
End-to-end text-to-speech synthesis (TTS), which generates speech sounds directly from strings of texts or phonemes, has improved the quality of speech synthesis over the conventional TTS. However, most previous studies have been evaluated based on subjective naturalness and have not objectively examined whether they can reproduce pitch patterns of phonological phenomena such as downstep, rhythmic boost, and initial lowering that reflect syntactic structures in Japanese. These phenomena can be linguistically explained by phonological constraints and the syntax$\unicode{x2013}$prosody mapping h…
Sequence-based neural networks show significant sensitivity to syntactic structure, but they still perform less well on syntactic tasks than tree-based networks. Such tree-based networks can be provided with a constituency parse, a dependency parse, or both. We evaluate which of these two representational schemes more effectively introduces biases for syntactic structure that increase performance on the subject-verb agreement prediction task. We find that a constituency-based network generalizes more robustly than a dependency-based one, and that combining the two types of structure does not y…
Particles fullfill several distinct central roles in the Japanese language. They can mark arguments as well as adjuncts, can be functional or have semantic funtions. There is, however, no straightforward matching from particles to functions, as, e.g., GA can mark the subject, the object or an adjunct of a sentence. Particles can cooccur. Verbal arguments that could be identified by particles can be eliminated in the Japanese sentence. And finally, in spoken language particles are often omitted. A proper treatment of particles is thus necessary to make an analysis of Japanese sentences possible…
Causal Interventions Reveal Typologically Organized Syntactic Mechanisms in Multilingual Language Models
The University of Texas at Austin（德克萨斯大学奥斯汀分校）；Saarland University（萨尔大学）；Leipzig University（莱比锡大学）；Heidelberg Institute for Theoretical Studies（海德堡理论研究所）
Bar Ilan University（巴伊兰大学）；Allen Institute for Artificial Intelligence（艾伦人工智能研究所）；Johns Hopkins University（约翰斯·霍普金斯大学）
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 langu…
With a growing interest in modeling inherent subjectivity in natural language, we present a linguistically-motivated process to understand and analyze the writing style of individuals from three perspectives: lexical, syntactic, and semantic. We discuss the stylistically expressive elements within each of these levels and use existing methods to quantify the linguistic intuitions related to some of these elements. We show that such a multi-level analysis is useful for developing a well-knit understanding of style - which is independent of the natural language task at hand, and also demonstrate…
Liu et al. (2017) provide a comprehensive account of research on dependency distance in human languages. While the article is a very rich and useful report on this complex subject, here I will expand on a few specific issues where research in computational linguistics (specifically natural language processing) can inform DDM research, and vice versa. These aspects have not been explored much in the article by Liu et al. or elsewhere, probably due to the little overlap between both research communities, but they may provide interesting insights for improving our understanding of the evolution o…
Beyond Memorization: Assessing Semantic Generalization in Large Language Models Using Phrasal Constructions
Georgetown University（乔治城大学）；University of Bath（巴斯大学）；DEVCOM U.S. Army Research Laboratory（美国陆军研究实验室）；University of Maryland, College Park（马里兰大学学院公园分校）
Many recent works have demonstrated that unsupervised sentence representations of neural networks encode syntactic information by observing that neural language models are able to predict the agreement between a verb and its subject. We take a critical look at this line of research by showing that it is possible to achieve high accuracy on this agreement task with simple surface heuristics, indicating a possible flaw in our assessment of neural networks' syntactic ability. Our fine-grained analyses of results on the long-range French object-verb agreement show that contrary to LSTMs, Transform…
Prior work has shown that structural supervision helps English language models learn generalizations about syntactic phenomena such as subject-verb agreement. However, it remains unclear if such an inductive bias would also improve language models' ability to learn grammatical dependencies in typologically different languages. Here we investigate this question in Mandarin Chinese, which has a logographic, largely syllable-based writing system; different word order; and sparser morphology than English. We train LSTMs, Recurrent Neural Network Grammars, Transformer language models, and Transform…
