The success of pre-trained contextualized representations has prompted researchers to analyze the... more The success of pre-trained contextualized representations has prompted researchers to analyze them for the presence of linguistic information. Indeed, it is natural to assume that these pre-trained representations do encode some level of linguistic knowledge as they have brought about large empirical improvements on a wide variety of NLP tasks, which suggests they are learning true linguistic generalization. In this work, we focus on intrinsic probing, an analysis technique where the goal is not only to identify whether a representation encodes a linguistic attribute, but also to pinpoint where this attribute is encoded. We propose a novel latent-variable formulation for constructing intrinsic probes and derive a tractable variational approximation to the log-likelihood. Our results show that our model is versatile and yields tighter mutual information estimates than two intrinsic probes previously proposed in the literature. Finally, we find empirical evidence that pre-trained repr...
Ethics is one of the longest standing intellectual endeavors of humanity. In recent years, the fi... more Ethics is one of the longest standing intellectual endeavors of humanity. In recent years, the fields of AI and NLP have attempted to wrangle with how learning systems that interact with humans should be constrained to behave ethically. One proposal in this vein is the construction of morality models that can take in arbitrary text and output a moral judgment about the situation described. In this work, we focus on a single case study of the recently proposed Delphi model and offer a critique of the project’s proposed method of automating morality judgments. Through an audit of Delphi, we examine broader issues that would be applicable to any similar attempt. We conclude with a discussion of how machine ethics could usefully proceed, by focusing on current and near-future uses of technology, in a way that centers around transparency, democratic values, and allows for straightforward accountability.
Natural Language Understanding has witnessed a watershed moment with the introduction of large pr... more Natural Language Understanding has witnessed a watershed moment with the introduction of large pre-trained Transformer networks. These models achieve state-of-the-art on various tasks, notably including Natural Language Inference (NLI). Many studies have shown that the large representation space imbibed by the models encodes some syntactic and semantic information. However, to really "know syntax", a model must recognize when its input violates syntactic rules and calculate inferences accordingly. In this work, we find that state-of-the-art NLI models, such as RoBERTa and BART are invariant to, and sometimes even perform better on, examples with randomly reordered words. With iterative search, we are able to construct randomized versions of NLI test sets, which contain permuted hypothesis-premise pairs with the same words as the original, yet are classified with perfect accuracy by large pre-trained models, as well as pre-Transformer state-of-the-art encoders. We find the ...
Weak definites in English have been widely studied as an example of when the<br> definite a... more Weak definites in English have been widely studied as an example of when the<br> definite article doesn't contribute uniqueness (Aguilar-Guevara & Zwarts 2011;<br> Aguilar-Guevara et al. 2014, among others). I take uniqueness to stem from the<br> interaction between definiteness and number within the noun phrase. From this<br> perspective, weak definites should be seen as a data point situated in the larger<br> cross-literature on number. One particular phenomenon from the literature on<br> number, the understudied class of the English bare institutional singulars (BISs),<br> has been discovered to share several semantic properties with weak definiteness,<br> namely number neutrality, referential deficiency, and lexical idiosyncrasy. In this<br> chapter, I postulate a shared account of English weak definites and BISs that uti-<br> lizes semantic root ambiguity (Rappaport Hovav & Levin 1998; Levinson 2014) as<br> a...
Verbs occur in different syntactic environments, or frames. We investigate whether artificial neu... more Verbs occur in different syntactic environments, or frames. We investigate whether artificial neural networks encode grammatical distinctions necessary for inferring the idiosyncratic frame-selectional properties of verbs. We introduce five datasets, collectively called FAVA, containing in aggregate nearly 10k sentences labeled for grammatical acceptability, illustrating different verbal argument structure alternations. We then test whether models can distinguish acceptable English verb-frame combinations from unacceptable ones using a sentence embedding alone. For converging evidence, we further construct LaVA, a corresponding word-level dataset, and investigate whether the same syntactic features can be extracted from word embeddings. Our models perform reliable classifications for some verbal alternations but not others, suggesting that while these representations do encode fine-grained lexical information, it is incomplete or can be hard to extract. Further, differences between ...
This paper presents the results of the RepEval 2017 Shared Task, which evaluated neural network s... more This paper presents the results of the RepEval 2017 Shared Task, which evaluated neural network sentence representation learning models on the Multi-Genre Natural Language Inference corpus (MultiNLI) recently introduced by Williams et al. (2017). All of the five participating teams beat the bidirectional LSTM (BiLSTM) and continuous bag of words baselines reported in Williams et al.. The best single model used stacked BiLSTMs with residual connections to extract sentence features and reached 74.5% accuracy on the genre-matched test set. Surprisingly, the results of the competition were fairly consistent across the genre-matched and genre-mismatched test sets, and across subsets of the test data representing a variety of linguistic phenomena, suggesting that all of the submitted systems learned reasonably domain-independent representations for sentence meaning.
This paper introduces the Multi-Genre Natural Language Inference (MultiNLI) corpus, a dataset des... more This paper introduces the Multi-Genre Natural Language Inference (MultiNLI) corpus, a dataset designed for use in the development and evaluation of machine learning models for sentence understanding. In addition to being one of the largest corpora available for the task of NLI, at 433k examples, this corpus improves upon available resources in its coverage: it offers data from ten distinct genres of written and spoken English--making it possible to evaluate systems on nearly the full complexity of the language--and it offers an explicit setting for the evaluation of cross-genre domain adaptation.
A possible explanation for the impressive performance of masked language model (MLM) pre-training... more A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks almost entirely due to their ability to model higher-order word co-occurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and show that these models still achieve high accuracy after fine-tuning on many downstream tasks -- including on tasks specifically designed to be challenging for models that ignore word order. Our models perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pre-training, and underscore the importance of curating challe...
While idiosyncrasies of the Chinese classifier system have been a richly studied topic among ling... more While idiosyncrasies of the Chinese classifier system have been a richly studied topic among linguists (Adams and Conklin, 1973; Erbaugh, 1986; Lakoff, 1986), not much work has been done to quantify them with statistical methods. In this paper, we introduce an information-theoretic approach to measuring idiosyncrasy; we examine how much the uncertainty in Mandarin Chinese classifiers can be reduced by knowing semantic information about the nouns that the classifiers modify. Using the empirical distribution of classifiers from the parsed Chinese Gigaword corpus (Graff et al., 2005), we compute the mutual information (in bits) between the distribution over classifiers and distributions over other linguistic quantities. We investigate whether semantic classes of nouns and adjectives differ in how much they reduce uncertainty in classifier choice, and find that it is not fully idiosyncratic; while there are no obvious trends for the majority of semantic classes, shape nouns reduce uncer...
Recent work on the problem of latent tree learning has made it possible to train neural networks ... more Recent work on the problem of latent tree learning has made it possible to train neural networks that learn to both parse a sentence and use the resulting parse to interpret the sentence, all without exposure to ground-truth parse trees at training time. Surprisingly, these models often perform better at sentence understanding tasks than models that use parse trees from conventional parsers. This paper aims to investigate what these latent tree learning models learn. We replicate two such models in a shared codebase and find that (i) only one of these models outperforms conventional tree-structured models on sentence classification, (ii) its parsing strategies are not especially consistent across random restarts, (iii) the parses it produces tend to be shallower than standard Penn Treebank (PTB) parses, and (iv) they do not resemble those of PTB or any other semantic or syntactic formalism that the authors are aware of.
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
Recent investigations into the inner-workings of state-of-the-art large-scale pre-trained Transfo... more Recent investigations into the inner-workings of state-of-the-art large-scale pre-trained Transformer-based Natural Language Understanding (NLU) models indicate that they appear to know humanlike syntax, at least to some extent. We provide novel evidence that complicates this claim: we find that state-of-the-art Natural Language Inference (NLI) models assign the same labels to permuted examples as they do to the original, i.e. they are largely invariant to random wordorder permutations. This behavior notably differs from that of humans; we struggle with ungrammatical sentences. To measure the severity of this issue, we propose a suite of metrics and investigate which properties of particular permutations lead models to be word-order invariant. In the MNLI dataset, for example, we find almost all (98.7%) examples contain at least one permutation which elicits the gold label. Models are sometimes even able to assign gold labels to permutations that they originally failed to predict correctly. We provide a comprehensive empirical evaluation of this phenomenon, and further show that this issue exists for both Transformers and pre-Transformer RNN / ConvNet based encoders, as well as across multiple languages (English and Mandarin Chinese). Our code and data are available at https://github.com/facebookresearch/unlu.
Proceedings of the 25th Conference on Computational Natural Language Learning
Inflectional morphology has since long been a useful testing ground for broader questions about g... more Inflectional morphology has since long been a useful testing ground for broader questions about generalisation in language and the viability of neural network models as cognitive models of language. Here, in line with that tradition, we explore how recurrent neural networks acquire the complex German plural system and reflect upon how their strategy compares to human generalisation and rule-based models of this system. We perform analyses including behavioural experiments, diagnostic classification, representation analysis and causal interventions, suggesting that the models rely on features that are also key predictors in rule-based models of German plurals. However, the models also display shortcut learning, which is crucial to overcome in search of more cognitively plausible generalisation behaviour.
Proceedings of the Fourth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP
Given the increasingly prominent role NLP models (will) play in our lives, it is important for hu... more Given the increasingly prominent role NLP models (will) play in our lives, it is important for human expectations of model behavior to align with actual model behavior. Using Natural Language Inference (NLI) as a case study, we investigate the extent to which human-generated explanations of models' inference decisions align with how models actually make these decisions. More specifically, we define three alignment metrics that quantify how well natural language explanations align with model sensitivity to input words, as measured by integrated gradients. Then, we evaluate eight different models (the base and large versions of BERT, RoBERTa and ELECTRA, as well as an RNN and bag-of-words model), and find that the BERT-base model has the highest alignment with human-generated explanations, for all alignment metrics. Focusing in on transformers, we find that the base versions tend to have higher alignment with human-generated explanations than their larger counterparts, suggesting that increasing the number of model parameters leads, in some cases, to worse alignment with human explanations. Finally, we find that a model's alignment with human explanations is not predicted by the model's accuracy, suggesting that accuracy and alignment are complementary ways to evaluate models. * The work was conducted during an internship at Facebook AI Research.
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
A possible explanation for the impressive performance of masked language model (MLM) pre-training... more A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks mostly due to their ability to model higher-order word cooccurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and we show that these models still achieve high accuracy after finetuning on many downstream tasks-including tasks specifically designed to be challenging for models that ignore word order. Our models also perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pretraining, and they underscore the importance of curating challenging evaluation datasets that require deeper linguistic knowledge.
Weak definites in English have been widely studied as an example of when the definite article doe... more Weak definites in English have been widely studied as an example of when the definite article doesn’t contribute uniqueness (Aguilar-Guevara & Zwarts 2011; Aguilar-Guevara et al. 2014, among others). I take uniqueness to stem from the interaction between definiteness and number within the noun phrase. From this perspective, weak definites should be seen as a data point situated in the larger cross-literature on number. One particular phenomenon from the literature on number, the understudied class of the English bare institutional singulars (BISs), has been discovered to share several semantic properties with weak definiteness, namely number neutrality, referential deficiency, and lexical idiosyncrasy. In this chapter, I postulate a shared account of English weak definites and BISs that utilizes semantic root ambiguity (Rappaport Hovav & Levin 1998; Levinson 2014) as a way to account for these facts. This account has syntactic consequences that resonate with recent morphosyntactic a...
Recent work on reinforcement learning and other gradient estimators for latent tree learning has ... more Recent work on reinforcement learning and other gradient estimators for latent tree learning has made it possible to train neural networks that learn to both parse a sentence and use the resulting parse to interpret the sentence, all without exposure to ground-truth parse trees at training time. Surprisingly, these models often perform better at sentence understanding tasks than models that use parse trees from conventional parsers. This paper aims to investigate what these latent tree learning models learn. We replicate two such models in a shared codebase and find that (i) they do outperform baselines on sentence classification, but that (ii) their parsing strategies are not especially consistent across random restarts, (iii) the parses they produce tend to be shallower than PTB parses, and (iv) these do not resemble those of PTB or of any other recognizable semantic or syntactic grammar formalism.
A possible explanation for the impressive performance of masked language model (MLM) pre-training... more A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks almost entirely due to their ability to model higher-order word co-occurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and show that these models still achieve high accuracy after fine-tuning on many downstream tasks— including on tasks specifically designed to be challenging for models that ignore word order. Our models perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pretraining, and underscore the importance of curating challengi...
The success of pre-trained contextualized representations has prompted researchers to analyze the... more The success of pre-trained contextualized representations has prompted researchers to analyze them for the presence of linguistic information. Indeed, it is natural to assume that these pre-trained representations do encode some level of linguistic knowledge as they have brought about large empirical improvements on a wide variety of NLP tasks, which suggests they are learning true linguistic generalization. In this work, we focus on intrinsic probing, an analysis technique where the goal is not only to identify whether a representation encodes a linguistic attribute, but also to pinpoint where this attribute is encoded. We propose a novel latent-variable formulation for constructing intrinsic probes and derive a tractable variational approximation to the log-likelihood. Our results show that our model is versatile and yields tighter mutual information estimates than two intrinsic probes previously proposed in the literature. Finally, we find empirical evidence that pre-trained repr...
Ethics is one of the longest standing intellectual endeavors of humanity. In recent years, the fi... more Ethics is one of the longest standing intellectual endeavors of humanity. In recent years, the fields of AI and NLP have attempted to wrangle with how learning systems that interact with humans should be constrained to behave ethically. One proposal in this vein is the construction of morality models that can take in arbitrary text and output a moral judgment about the situation described. In this work, we focus on a single case study of the recently proposed Delphi model and offer a critique of the project’s proposed method of automating morality judgments. Through an audit of Delphi, we examine broader issues that would be applicable to any similar attempt. We conclude with a discussion of how machine ethics could usefully proceed, by focusing on current and near-future uses of technology, in a way that centers around transparency, democratic values, and allows for straightforward accountability.
Natural Language Understanding has witnessed a watershed moment with the introduction of large pr... more Natural Language Understanding has witnessed a watershed moment with the introduction of large pre-trained Transformer networks. These models achieve state-of-the-art on various tasks, notably including Natural Language Inference (NLI). Many studies have shown that the large representation space imbibed by the models encodes some syntactic and semantic information. However, to really "know syntax", a model must recognize when its input violates syntactic rules and calculate inferences accordingly. In this work, we find that state-of-the-art NLI models, such as RoBERTa and BART are invariant to, and sometimes even perform better on, examples with randomly reordered words. With iterative search, we are able to construct randomized versions of NLI test sets, which contain permuted hypothesis-premise pairs with the same words as the original, yet are classified with perfect accuracy by large pre-trained models, as well as pre-Transformer state-of-the-art encoders. We find the ...
Weak definites in English have been widely studied as an example of when the<br> definite a... more Weak definites in English have been widely studied as an example of when the<br> definite article doesn't contribute uniqueness (Aguilar-Guevara & Zwarts 2011;<br> Aguilar-Guevara et al. 2014, among others). I take uniqueness to stem from the<br> interaction between definiteness and number within the noun phrase. From this<br> perspective, weak definites should be seen as a data point situated in the larger<br> cross-literature on number. One particular phenomenon from the literature on<br> number, the understudied class of the English bare institutional singulars (BISs),<br> has been discovered to share several semantic properties with weak definiteness,<br> namely number neutrality, referential deficiency, and lexical idiosyncrasy. In this<br> chapter, I postulate a shared account of English weak definites and BISs that uti-<br> lizes semantic root ambiguity (Rappaport Hovav & Levin 1998; Levinson 2014) as<br> a...
Verbs occur in different syntactic environments, or frames. We investigate whether artificial neu... more Verbs occur in different syntactic environments, or frames. We investigate whether artificial neural networks encode grammatical distinctions necessary for inferring the idiosyncratic frame-selectional properties of verbs. We introduce five datasets, collectively called FAVA, containing in aggregate nearly 10k sentences labeled for grammatical acceptability, illustrating different verbal argument structure alternations. We then test whether models can distinguish acceptable English verb-frame combinations from unacceptable ones using a sentence embedding alone. For converging evidence, we further construct LaVA, a corresponding word-level dataset, and investigate whether the same syntactic features can be extracted from word embeddings. Our models perform reliable classifications for some verbal alternations but not others, suggesting that while these representations do encode fine-grained lexical information, it is incomplete or can be hard to extract. Further, differences between ...
This paper presents the results of the RepEval 2017 Shared Task, which evaluated neural network s... more This paper presents the results of the RepEval 2017 Shared Task, which evaluated neural network sentence representation learning models on the Multi-Genre Natural Language Inference corpus (MultiNLI) recently introduced by Williams et al. (2017). All of the five participating teams beat the bidirectional LSTM (BiLSTM) and continuous bag of words baselines reported in Williams et al.. The best single model used stacked BiLSTMs with residual connections to extract sentence features and reached 74.5% accuracy on the genre-matched test set. Surprisingly, the results of the competition were fairly consistent across the genre-matched and genre-mismatched test sets, and across subsets of the test data representing a variety of linguistic phenomena, suggesting that all of the submitted systems learned reasonably domain-independent representations for sentence meaning.
This paper introduces the Multi-Genre Natural Language Inference (MultiNLI) corpus, a dataset des... more This paper introduces the Multi-Genre Natural Language Inference (MultiNLI) corpus, a dataset designed for use in the development and evaluation of machine learning models for sentence understanding. In addition to being one of the largest corpora available for the task of NLI, at 433k examples, this corpus improves upon available resources in its coverage: it offers data from ten distinct genres of written and spoken English--making it possible to evaluate systems on nearly the full complexity of the language--and it offers an explicit setting for the evaluation of cross-genre domain adaptation.
A possible explanation for the impressive performance of masked language model (MLM) pre-training... more A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks almost entirely due to their ability to model higher-order word co-occurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and show that these models still achieve high accuracy after fine-tuning on many downstream tasks -- including on tasks specifically designed to be challenging for models that ignore word order. Our models perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pre-training, and underscore the importance of curating challe...
While idiosyncrasies of the Chinese classifier system have been a richly studied topic among ling... more While idiosyncrasies of the Chinese classifier system have been a richly studied topic among linguists (Adams and Conklin, 1973; Erbaugh, 1986; Lakoff, 1986), not much work has been done to quantify them with statistical methods. In this paper, we introduce an information-theoretic approach to measuring idiosyncrasy; we examine how much the uncertainty in Mandarin Chinese classifiers can be reduced by knowing semantic information about the nouns that the classifiers modify. Using the empirical distribution of classifiers from the parsed Chinese Gigaword corpus (Graff et al., 2005), we compute the mutual information (in bits) between the distribution over classifiers and distributions over other linguistic quantities. We investigate whether semantic classes of nouns and adjectives differ in how much they reduce uncertainty in classifier choice, and find that it is not fully idiosyncratic; while there are no obvious trends for the majority of semantic classes, shape nouns reduce uncer...
Recent work on the problem of latent tree learning has made it possible to train neural networks ... more Recent work on the problem of latent tree learning has made it possible to train neural networks that learn to both parse a sentence and use the resulting parse to interpret the sentence, all without exposure to ground-truth parse trees at training time. Surprisingly, these models often perform better at sentence understanding tasks than models that use parse trees from conventional parsers. This paper aims to investigate what these latent tree learning models learn. We replicate two such models in a shared codebase and find that (i) only one of these models outperforms conventional tree-structured models on sentence classification, (ii) its parsing strategies are not especially consistent across random restarts, (iii) the parses it produces tend to be shallower than standard Penn Treebank (PTB) parses, and (iv) they do not resemble those of PTB or any other semantic or syntactic formalism that the authors are aware of.
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
Recent investigations into the inner-workings of state-of-the-art large-scale pre-trained Transfo... more Recent investigations into the inner-workings of state-of-the-art large-scale pre-trained Transformer-based Natural Language Understanding (NLU) models indicate that they appear to know humanlike syntax, at least to some extent. We provide novel evidence that complicates this claim: we find that state-of-the-art Natural Language Inference (NLI) models assign the same labels to permuted examples as they do to the original, i.e. they are largely invariant to random wordorder permutations. This behavior notably differs from that of humans; we struggle with ungrammatical sentences. To measure the severity of this issue, we propose a suite of metrics and investigate which properties of particular permutations lead models to be word-order invariant. In the MNLI dataset, for example, we find almost all (98.7%) examples contain at least one permutation which elicits the gold label. Models are sometimes even able to assign gold labels to permutations that they originally failed to predict correctly. We provide a comprehensive empirical evaluation of this phenomenon, and further show that this issue exists for both Transformers and pre-Transformer RNN / ConvNet based encoders, as well as across multiple languages (English and Mandarin Chinese). Our code and data are available at https://github.com/facebookresearch/unlu.
Proceedings of the 25th Conference on Computational Natural Language Learning
Inflectional morphology has since long been a useful testing ground for broader questions about g... more Inflectional morphology has since long been a useful testing ground for broader questions about generalisation in language and the viability of neural network models as cognitive models of language. Here, in line with that tradition, we explore how recurrent neural networks acquire the complex German plural system and reflect upon how their strategy compares to human generalisation and rule-based models of this system. We perform analyses including behavioural experiments, diagnostic classification, representation analysis and causal interventions, suggesting that the models rely on features that are also key predictors in rule-based models of German plurals. However, the models also display shortcut learning, which is crucial to overcome in search of more cognitively plausible generalisation behaviour.
Proceedings of the Fourth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP
Given the increasingly prominent role NLP models (will) play in our lives, it is important for hu... more Given the increasingly prominent role NLP models (will) play in our lives, it is important for human expectations of model behavior to align with actual model behavior. Using Natural Language Inference (NLI) as a case study, we investigate the extent to which human-generated explanations of models' inference decisions align with how models actually make these decisions. More specifically, we define three alignment metrics that quantify how well natural language explanations align with model sensitivity to input words, as measured by integrated gradients. Then, we evaluate eight different models (the base and large versions of BERT, RoBERTa and ELECTRA, as well as an RNN and bag-of-words model), and find that the BERT-base model has the highest alignment with human-generated explanations, for all alignment metrics. Focusing in on transformers, we find that the base versions tend to have higher alignment with human-generated explanations than their larger counterparts, suggesting that increasing the number of model parameters leads, in some cases, to worse alignment with human explanations. Finally, we find that a model's alignment with human explanations is not predicted by the model's accuracy, suggesting that accuracy and alignment are complementary ways to evaluate models. * The work was conducted during an internship at Facebook AI Research.
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
A possible explanation for the impressive performance of masked language model (MLM) pre-training... more A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks mostly due to their ability to model higher-order word cooccurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and we show that these models still achieve high accuracy after finetuning on many downstream tasks-including tasks specifically designed to be challenging for models that ignore word order. Our models also perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pretraining, and they underscore the importance of curating challenging evaluation datasets that require deeper linguistic knowledge.
Weak definites in English have been widely studied as an example of when the definite article doe... more Weak definites in English have been widely studied as an example of when the definite article doesn’t contribute uniqueness (Aguilar-Guevara & Zwarts 2011; Aguilar-Guevara et al. 2014, among others). I take uniqueness to stem from the interaction between definiteness and number within the noun phrase. From this perspective, weak definites should be seen as a data point situated in the larger cross-literature on number. One particular phenomenon from the literature on number, the understudied class of the English bare institutional singulars (BISs), has been discovered to share several semantic properties with weak definiteness, namely number neutrality, referential deficiency, and lexical idiosyncrasy. In this chapter, I postulate a shared account of English weak definites and BISs that utilizes semantic root ambiguity (Rappaport Hovav & Levin 1998; Levinson 2014) as a way to account for these facts. This account has syntactic consequences that resonate with recent morphosyntactic a...
Recent work on reinforcement learning and other gradient estimators for latent tree learning has ... more Recent work on reinforcement learning and other gradient estimators for latent tree learning has made it possible to train neural networks that learn to both parse a sentence and use the resulting parse to interpret the sentence, all without exposure to ground-truth parse trees at training time. Surprisingly, these models often perform better at sentence understanding tasks than models that use parse trees from conventional parsers. This paper aims to investigate what these latent tree learning models learn. We replicate two such models in a shared codebase and find that (i) they do outperform baselines on sentence classification, but that (ii) their parsing strategies are not especially consistent across random restarts, (iii) the parses they produce tend to be shallower than PTB parses, and (iv) these do not resemble those of PTB or of any other recognizable semantic or syntactic grammar formalism.
A possible explanation for the impressive performance of masked language model (MLM) pre-training... more A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks almost entirely due to their ability to model higher-order word co-occurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and show that these models still achieve high accuracy after fine-tuning on many downstream tasks— including on tasks specifically designed to be challenging for models that ignore word order. Our models perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pretraining, and underscore the importance of curating challengi...
Transformer based models are the modern work horses for neural machine translation (NMT), reachin... more Transformer based models are the modern work horses for neural machine translation (NMT), reaching state of the art across several benchmarks. Despite their impressive accuracy, we observe a systemic and rudimentary class of errors made by transformer based models with regards to translating from a language that doesn’t mark gender on nouns into others that do. We find that even when the surrounding context provides unambiguous evidence of the appropriate grammatical gender marking, no transformer based model we tested was able to accurately gender occupation nouns systematically. We release an evaluation scheme and dataset for measuring the ability of transformer based NMT models to translate gender morphology correctly in unambiguous contexts across syntactically diverse sentences. Our dataset translates from an English source into 20 languages from several different language families. With the availability of this dataset, our hope is that the NMT community can iterate on solutio...
Uploads
Talks by Adina Williams
Papers by Adina Williams