EMNLP 2020 论文列表

Deep transformer models have pushed performance on NLP tasks to new limits, suggesting sophisticated treatment of complex linguistic inputs, such as phrases. However, we have limited understanding of how these models handle representation of phrases, and whether this reflects sophisticated composition of phrase meaning like that done by humans. In this paper, we present systematic analysis of phrasal representations in state-of-the-art pre-trained transformers. We use tests leveraging human judgments of phrase similarity and meaning shift, and compare results before and after control of word overlap, to tease apart lexical effects versus composition effects. We find that phrase representation in these models relies heavily on word content, with little evidence of nuanced composition. We also identify variations in phrase representation quality across models, layers, and representation types, and make corresponding recommendations for usage of representations from these models.

Span identification (in short, span ID) tasks such as chunking, NER, or code-switching detection, ask models to identify and classify relevant spans in a text. Despite being a staple of NLP, and sharing a common structure, there is little insight on how these tasks’ properties influence their difficulty, and thus little guidance on what model families work well on span ID tasks, and why. We analyze span ID tasks via performance prediction, estimating how well neural architectures do on different tasks. Our contributions are: (a) we identify key properties of span ID tasks that can inform performance prediction; (b) we carry out a large-scale experiment on English data, building a model to predict performance for unseen span ID tasks that can support architecture choices; (c), we investigate the parameters of the meta model, yielding new insights on how model and task properties interact to affect span ID performance. We find, e.g., that span frequency is especially important for LSTMs, and that CRFs help when spans are infrequent and boundaries non-distinctive.

While a lot of analysis has been carried to demonstrate linguistic knowledge captured by the representations learned within deep NLP models, very little attention has been paid towards individual neurons. We carry outa neuron-level analysis using core linguistic tasks of predicting morphology, syntax and semantics, on pre-trained language models, with questions like: i) do individual neurons in pre-trained models capture linguistic information? ii) which parts of the network learn more about certain linguistic phenomena? iii) how distributed or focused is the information? and iv) how do various architectures differ in learning these properties? We found small subsets of neurons to predict linguistic tasks, with lower level tasks (such as morphology) localized in fewer neurons, compared to higher level task of predicting syntax. Our study also reveals interesting cross architectural comparisons. For example, we found neurons in XLNet to be more localized and disjoint when predicting properties compared to BERT and others, where they are more distributed and coupled.

Pre-training large language models has become a standard in the natural language processing community. Such models are pre-trained on generic data (e.g. BookCorpus and English Wikipedia) and often fine-tuned on tasks in the same domain. However, in order to achieve state-of-the-art performance on out of domain tasks such as clinical named entity recognition and relation extraction, additional in domain pre-training is required. In practice, staged multi-domain pre-training presents performance deterioration in the form of catastrophic forgetting (CF) when evaluated on a generic benchmark such as GLUE. In this paper we conduct an empirical investigation into known methods to mitigate CF. We find that elastic weight consolidation provides best overall scores yielding only a 0.33% drop in performance across seven generic tasks while remaining competitive in bio-medical tasks. Furthermore, we explore gradient and latent clustering based data selection techniques to improve coverage when using elastic weight consolidation and experience replay methods.

Benchmarks such as GLUE have helped drive advances in NLP by incentivizing the creation of more accurate models. While this leaderboard paradigm has been remarkably successful, a historical focus on performance-based evaluation has been at the expense of other qualities that the NLP community values in models, such as compactness, fairness, and energy efficiency. In this opinion paper, we study the divergence between what is incentivized by leaderboards and what is useful in practice through the lens of microeconomic theory. We frame both the leaderboard and NLP practitioners as consumers and the benefit they get from a model as its utility to them. With this framing, we formalize how leaderboards – in their current form – can be poor proxies for the NLP community at large. For example, a highly inefficient model would provide less utility to practitioners but not to a leaderboard, since it is a cost that only the former must bear. To allow practitioners to better estimate a model’s utility to them, we advocate for more transparency on leaderboards, such as the reporting of statistics that are of practical concern (e.g., model size, energy efficiency, and inference latency).

The deep inside-outside recursive autoencoder (DIORA; Drozdov et al. 2019) is a self-supervised neural model that learns to induce syntactic tree structures for input sentences without access to labeled training data. In this paper, we discover that while DIORA exhaustively encodes all possible binary trees of a sentence with a soft dynamic program, its vector averaging approach is locally greedy and cannot recover from errors when computing the highest scoring parse tree in bottom-up chart parsing. To fix this issue, we introduce S-DIORA, an improved variant of DIORA that encodes a single tree rather than a softly-weighted mixture of trees by employing a hard argmax operation and a beam at each cell in the chart. Our experiments show that through fine-tuning a pre-trained DIORA with our new algorithm, we improve the state of the art in unsupervised constituency parsing on the English WSJ Penn Treebank by 2.2-6% F1, depending on the data used for fine-tuning.

We describe a fully unsupervised cross-lingual transfer approach for part-of-speech (POS) tagging under a truly low resource scenario. We assume access to parallel translations between the target language and one or more source languages for which POS taggers are available. We use the Bible as parallel data in our experiments: small size, out-of-domain and covering many diverse languages. Our approach innovates in three ways: 1) a robust approach of selecting training instances via cross-lingual annotation projection that exploits best practices of unsupervised type and token constraints, word-alignment confidence and density of projected POS, 2) a Bi-LSTM architecture that uses contextualized word embeddings, affix embeddings and hierarchical Brown clusters, and 3) an evaluation on 12 diverse languages in terms of language family and morphological typology. In spite of the use of limited and out-of-domain parallel data, our experiments demonstrate significant improvements in accuracy over previous work. In addition, we show that using multi-source information, either via projection or output combination, improves the performance for most target languages.

The connection between dependency trees and spanning trees is exploited by the NLP community to train and to decode graph-based dependency parsers. However, the NLP literature has missed an important difference between the two structures: only one edge may emanate from the root in a dependency tree. We analyzed the output of state-of-the-art parsers on many languages from the Universal Dependency Treebank: although these parsers are often able to learn that trees which violate the constraint should be assigned lower probabilities, their ability to do so unsurprisingly de-grades as the size of the training set decreases. In fact, the worst constraint-violation rate we observe is 24%. Prior work has proposed an inefficient algorithm to enforce the constraint, which adds a factor of n to the decoding runtime. We adapt an algorithm due to Gabow and Tarjan (1984) to dependency parsing, which satisfies the constraint without compromising the original runtime.

We propose a method for unsupervised parsing based on the linguistic notion of a constituency test. One type of constituency test involves modifying the sentence via some transformation (e.g. replacing the span with a pronoun) and then judging the result (e.g. checking if it is grammatical). Motivated by this idea, we design an unsupervised parser by specifying a set of transformations and using an unsupervised neural acceptability model to make grammaticality decisions. To produce a tree given a sentence, we score each span by aggregating its constituency test judgments, and we choose the binary tree with the highest total score. While this approach already achieves performance in the range of current methods, we further improve accuracy by fine-tuning the grammaticality model through a refinement procedure, where we alternate between improving the estimated trees and improving the grammaticality model. The refined model achieves 62.8 F1 on the Penn Treebank test set, an absolute improvement of 7.6 points over the previously best published result.

Morphologically rich languages seem to benefit from joint processing of morphology and syntax, as compared to pipeline architectures. We propose a graph-based model for joint morphological parsing and dependency parsing in Sanskrit. Here, we extend the Energy based model framework (Krishna et al., 2020), proposed for several structured prediction tasks in Sanskrit, in 2 simple yet significant ways. First, the framework’s default input graph generation method is modified to generate a multigraph, which enables the use of an exact search inference. Second, we prune the input search space using a linguistically motivated approach, rooted in the traditional grammatical analysis of Sanskrit. Our experiments show that the morphological parsing from our joint model outperforms standalone morphological parsers. We report state of the art results in morphological parsing, and in dependency parsing, both in standalone (with gold morphological tags) and joint morphosyntactic parsing setting.

Exposure to violent, sexual, or substance-abuse content in media increases the willingness of children and adolescents to imitate similar behaviors. Computational methods that identify portrayals of risk behaviors from audio-visual cues are limited in their applicability to films in post-production, where modifications might be prohibitively expensive. To address this limitation, we propose a model that estimates content ratings based on the language use in movie scripts, making our solution available at the earlier stages of creative production. Our model significantly improves the state-of-the-art by adapting novel techniques to learn better movie representations from the semantic and sentiment aspects of a character’s language use, and by leveraging the co-occurrence of risk behaviors, following a multi-task approach. Additionally, we show how this approach can be useful to learn novel insights on the joint portrayal of these behaviors, and on the subtleties that filmmakers may otherwise not pick up on.

The music genre perception expressed through human annotations of artists or albums varies significantly across language-bound cultures. These variations cannot be modeled as mere translations since we also need to account for cultural differences in the music genre perception. In this work, we study the feasibility of obtaining relevant cross-lingual, culture-specific music genre annotations based only on language-specific semantic representations, namely distributed concept embeddings and ontologies. Our study, focused on six languages, shows that unsupervised cross-lingual music genre annotation is feasible with high accuracy, especially when combining both types of representations. This approach of studying music genres is the most extensive to date and has many implications in musicology and music information retrieval. Besides, we introduce a new, domain-dependent cross-lingual corpus to benchmark state of the art multilingual pre-trained embedding models.

We conduct a large scale empirical investigation of contextualized number prediction in running text. Specifically, we consider two tasks: (1)masked number prediction– predict-ing a missing numerical value within a sentence, and (2)numerical anomaly detection–detecting an errorful numeric value within a sentence. We experiment with novel combinations of contextual encoders and output distributions over the real number line. Specifically, we introduce a suite of output distribution parameterizations that incorporate latent variables to add expressivity and better fit the natural distribution of numeric values in running text, and combine them with both recur-rent and transformer-based encoder architectures. We evaluate these models on two numeric datasets in the financial and scientific domain. Our findings show that output distributions that incorporate discrete latent variables and allow for multiple modes outperform simple flow-based counterparts on all datasets, yielding more accurate numerical pre-diction and anomaly detection. We also show that our models effectively utilize textual con-text and benefit from general-purpose unsupervised pretraining.

Word embedding models are typically able to capture the semantics of words via the distributional hypothesis, but fail to capture the numerical properties of numbers that appear in the text. This leads to problems with numerical reasoning involving tasks such as question answering. We propose a new methodology to assign and learn embeddings for numbers. Our approach creates Deterministic, Independent-of-Corpus Embeddings (the model is referred to as DICE) for numbers, such that their cosine similarity reflects the actual distance on the number line. DICE outperforms a wide range of pre-trained word embedding models across multiple examples of two tasks: (i) evaluating the ability to capture numeration and magnitude; and (ii) to perform list maximum, decoding, and addition. We further explore the utility of these embeddings in downstream tasks, by initializing numbers with our approach for the task of magnitude prediction. We also introduce a regularization approach to learn model-based embeddings of numbers in a contextual setting.

In this work, we present a new language pre-training model TNT (Text Normalization based pre-training of Transformers) for content moderation. Inspired by the masking strategy and text normalization, TNT is developed to learn language representation by training transformers to reconstruct text from four operation types typically seen in text manipulation: substitution, transposition, deletion, and insertion. Furthermore, the normalization involves the prediction of both operation types and token labels, enabling TNT to learn from more challenging tasks than the standard task of masked word recovery. As a result, the experiments demonstrate that TNT outperforms strong baselines on the hate speech classification task. Additional text normalization experiments and case studies show that TNT is a new potential approach to misspelling correction.

Multilingual pre-trained Transformers, such as mBERT (Devlin et al., 2019) and XLM-RoBERTa (Conneau et al., 2020a), have been shown to enable effective cross-lingual zero-shot transfer. However, their performance on Arabic information extraction (IE) tasks is not very well studied. In this paper, we pre-train a customized bilingual BERT, dubbed GigaBERT, that is designed specifically for Arabic NLP and English-to-Arabic zero-shot transfer learning. We study GigaBERT’s effectiveness on zero-short transfer across four IE tasks: named entity recognition, part-of-speech tagging, argument role labeling, and relation extraction. Our best model significantly outperforms mBERT, XLM-RoBERTa, and AraBERT (Antoun et al., 2020) in both the supervised and zero-shot transfer settings. We have made our pre-trained models publicly available at: https://github.com/lanwuwei/GigaBERT.

In this paper, we introduce an advanced Russian general language understanding evaluation benchmark – Russian SuperGLUE. Recent advances in the field of universal language models and transformers require the development of a methodology for their broad diagnostics and testing for general intellectual skills - detection of natural language inference, commonsense reasoning, ability to perform simple logical operations regardless of text subject or lexicon. For the first time, a benchmark of nine tasks, collected and organized analogically to the SuperGLUE methodology, was developed from scratch for the Russian language. We also provide baselines, human level evaluation, open-source framework for evaluating models, and an overall leaderboard of transformer models for the Russian language. Besides, we present the first results of comparing multilingual models in the translated diagnostic test set and offer the first steps to further expanding or assessing State-of-the-art models independently of language.

Document and discourse segmentation are two fundamental NLP tasks pertaining to breaking up text into constituents, which are commonly used to help downstream tasks such as information retrieval or text summarization. In this work, we propose three transformer-based architectures and provide comprehensive comparisons with previously proposed approaches on three standard datasets. We establish a new state-of-the-art, reducing in particular the error rates by a large margin in all cases. We further analyze model sizes and find that we can build models with many fewer parameters while keeping good performance, thus facilitating real-world applications.

There has been an influx of biomedical domain-specific language models, showing language models pre-trained on biomedical text perform better on biomedical domain benchmarks than those trained on general domain text corpora such as Wikipedia and Books. Yet, most works do not study the factors affecting each domain language application deeply. Additionally, the study of model size on domain-specific models has been mostly missing. We empirically study and evaluate several factors that can affect performance on domain language applications, such as the sub-word vocabulary set, model size, pre-training corpus, and domain transfer. We show consistent improvements on benchmarks with our larger BioMegatron model trained on a larger domain corpus, contributing to our understanding of domain language model applications. We demonstrate noticeable improvements over the previous state-of-the-art (SOTA) on standard biomedical NLP benchmarks of question answering, named entity recognition, and relation extraction. Code and checkpoints to reproduce our experiments are available at [github.com/NVIDIA/NeMo].