TACL 2021 论文列表

Phonological generalizations are finite-state. While Optimality Theory is a popular framework for modeling phonology, it is known to generate non-finite-state mappings and languages. This paper demonstrates that Optimality Theory is capable of generating non-context-free languages, contributing to the characterization of its generative capacity. This is achieved with minimal modification to the theory as it is standardly employed.

Recent approaches to data-to-text generation have adopted the very successful encoder-decoder architecture or variants thereof. These models generate text that is fluent (but often imprecise) and perform quite poorly at selecting appropriate content and ordering it coherently. To overcome some of these issues, we propose a neural model with a macro planning stage followed by a generation stage reminiscent of traditional methods which embrace separate modules for planning and surface realization. Macro plans represent high level organization of important content such as entities, events, and their interactions; they are learned from data and given as input to the generator. Extensive experiments on two data-to-text benchmarks (RotoWire and MLB) show that our approach outperforms competitive baselines in terms of automatic and human evaluation.

We introduce a novel paraphrastic augmentation strategy based on sentence-level lexically constrained paraphrasing and discriminative span alignment. Our approach allows for the large-scale expansion of existing datasets or the rapid creation of new datasets using a small, manually produced seed corpus. We demonstrate our approach with experiments on the Berkeley FrameNet Project, a large-scale language understanding effort spanning more than two decades of human labor. With four days of training data collection for a span alignment model and one day of parallel compute, we automatically generate and release to the community 495,300 unique (Frame,Trigger) pairs in diverse sentential contexts, a roughly 50-fold expansion atop FrameNet v1.7. The resulting dataset is intrinsically and extrinsically evaluated in detail, showing positive results on a downstream task.

Optical character recognition (OCR) is crucial for a deeper access to historical collections. OCR needs to account for orthographic variations, typefaces, or language evolution (i.e., new letters, word spellings), as the main source of character, word, or word segmentation transcription errors. For digital corpora of historical prints, the errors are further exacerbated due to low scan quality and lack of language standardization. For the task of OCR post-hoc correction, we propose a neural approach based on a combination of recurrent (RNN) and deep convolutional network (ConvNet) to correct OCR transcription errors. At character level we flexibly capture errors, and decode the corrected output based on a novel attention mechanism. Accounting for the input and output similarity, we propose a new loss function that rewards the model’s correcting behavior. Evaluation on a historical book corpus in German language shows that our models are robust in capturing diverse OCR transcription errors and reduce the word error rate of 32.3% by more than 89%.

Slang is a common type of informal language, but its flexible nature and paucity of data resources present challenges for existing natural language systems. We take an initial step toward machine generation of slang by developing a framework that models the speaker’s word choice in slang context. Our framework encodes novel slang meaning by relating the conventional and slang senses of a word while incorporating syntactic and contextual knowledge in slang usage. We construct the framework using a combination of probabilistic inference and neural contrastive learning. We perform rigorous evaluations on three slang dictionaries and show that our approach not only outperforms state-of-the-art language models, but also better predicts the historical emergence of slang word usages from 1960s to 2000s. We interpret the proposed models and find that the contrastively learned semantic space is sensitive to the similarities between slang and conventional senses of words. Our work creates opportunities for the automated generation and interpretation of informal language.

Models for question answering, dialogue agents, and summarization often interpret the meaning of a sentence in a rich context and use that meaning in a new context. Taking excerpts of text can be problematic, as key pieces may not be explicit in a local window. We isolate and define the problem of sentence decontextualization: taking a sentence together with its context and rewriting it to be interpretable out of context, while preserving its meaning. We describe an annotation procedure, collect data on the Wikipedia corpus, and use the data to train models to automatically decontextualize sentences. We present preliminary studies that show the value of sentence decontextualization in a user-facing task, and as preprocessing for systems that perform document understanding. We argue that decontextualization is an important subtask in many downstream applications, and that the definitions and resources provided can benefit tasks that operate on sentences that occur in a richer context.

The metrics standardly used to evaluate Natural Language Generation (NLG) models, such as BLEU or METEOR, fail to provide information on which linguistic factors impact performance. Focusing on Surface Realization (SR), the task of converting an unordered dependency tree into a well-formed sentence, we propose a framework for error analysis which permits identifying which features of the input affect the models’ results. This framework consists of two main components: (i) correlation analyses between a wide range of syntactic metrics and standard performance metrics and (ii) a set of techniques to automatically identify syntactic constructs that often co-occur with low performance scores. We demonstrate the advantages of our framework by performing error analysis on the results of 174 system runs submitted to the Multilingual SR shared tasks; we show that dependency edge accuracy correlate with automatic metrics thereby providing a more interpretable basis for evaluation; and we suggest ways in which our framework could be used to improve models and data. The framework is available in the form of a toolkit which can be used both by campaign organizers to provide detailed, linguistically interpretable feedback on the state of the art in multilingual SR, and by individual researchers to improve models and datasets.1

Most combinations of NLP tasks and language varieties lack in-domain examples for supervised training because of the paucity of annotated data. How can neural models make sample-efficient generalizations from task–language combinations with available data to low-resource ones? In this work, we propose a Bayesian generative model for the space of neural parameters. We assume that this space can be factorized into latent variables for each language and each task. We infer the posteriors over such latent variables based on data from seen task–language combinations through variational inference. This enables zero-shot classification on unseen combinations at prediction time. For instance, given training data for named entity recognition (NER) in Vietnamese and for part-of-speech (POS) tagging in Wolof, our model can perform accurate predictions for NER in Wolof. In particular, we experiment with a typologically diverse sample of 33 languages from 4 continents and 11 families, and show that our model yields comparable or better results than state-of-the-art, zero-shot cross-lingual transfer methods. Our code is available at github.com/cambridgeltl/parameter-factorization.

The scarcity of comprehensive up-to-date studies on evaluation metrics for text summarization and the lack of consensus regarding evaluation protocols continue to inhibit progress. We address the existing shortcomings of summarization evaluation methods along five dimensions: 1) we re-evaluate 14 automatic evaluation metrics in a comprehensive and consistent fashion using neural summarization model outputs along with expert and crowd-sourced human annotations; 2) we consistently benchmark 23 recent summarization models using the aforementioned automatic evaluation metrics; 3) we assemble the largest collection of summaries generated by models trained on the CNN/DailyMail news dataset and share it in a unified format; 4) we implement and share a toolkit that provides an extensible and unified API for evaluating summarization models across a broad range of automatic metrics; and 5) we assemble and share the largest and most diverse, in terms of model types, collection of human judgments of model-generated summaries on the CNN/Daily Mail dataset annotated by both expert judges and crowd-source workers. We hope that this work will help promote a more complete evaluation protocol for text summarization as well as advance research in developing evaluation metrics that better correlate with human judgments.

Task-oriented dialog (TOD) systems often need to formulate knowledge base (KB) queries corresponding to the user intent and use the query results to generate system responses. Existing approaches require dialog datasets to explicitly annotate these KB queries—these annotations can be time consuming, and expensive. In response, we define the novel problems of predicting the KB query and training the dialog agent, without explicit KB query annotation. For query prediction, we propose a reinforcement learning (RL) baseline, which rewards the generation of those queries whose KB results cover the entities mentioned in subsequent dialog. Further analysis reveals that correlation among query attributes in KB can significantly confuse memory augmented policy optimization (MAPO), an existing state of the art RL agent. To address this, we improve the MAPO baseline with simple but important modifications suited to our task. To train the full TOD system for our setting, we propose a pipelined approach: it independently predicts when to make a KB query (query position predictor), then predicts a KB query at the predicted position (query predictor), and uses the results of predicted query in subsequent dialog (next response predictor). Overall, our work proposes first solutions to our novel problem, and our analysis highlights the research challenges in training TOD systems without query annotation.

We present a language model that combines a large parametric neural network (i.e., a transformer) with a non-parametric episodic memory component in an integrated architecture. Our model uses extended short-term context by caching local hidden states—similar to transformer-XL—and global long-term memory by retrieving a set of nearest neighbor tokens at each timestep. We design a gating function to adaptively combine multiple information sources to make a prediction. This mechanism allows the model to use either local context, short-term memory, or long-term memory (or any combination of them) on an ad hoc basis depending on the context. Experiments on word-based and character-based language modeling datasets demonstrate the efficacy of our proposed method compared to strong baselines.

A key limitation in current datasets for multi-hop reasoning is that the required steps for answering the question are mentioned in it explicitly. In this work, we introduce StrategyQA, a question answering (QA) benchmark where the required reasoning steps are implicit in the question, and should be inferred using a strategy. A fundamental challenge in this setup is how to elicit such creative questions from crowdsourcing workers, while covering a broad range of potential strategies. We propose a data collection procedure that combines term-based priming to inspire annotators, careful control over the annotator population, and adversarial filtering for eliminating reasoning shortcuts. Moreover, we annotate each question with (1) a decomposition into reasoning steps for answering it, and (2) Wikipedia paragraphs that contain the answers to each step. Overall, StrategyQA includes 2,780 examples, each consisting of a strategy question, its decomposition, and evidence paragraphs. Analysis shows that questions in StrategyQA are short, topic-diverse, and cover a wide range of strategies. Empirically, we show that humans perform well (87%) on this task, while our best baseline reaches an accuracy of ∼ 66%.

Dual encoders perform retrieval by encoding documents and queries into dense low-dimensional vectors, scoring each document by its inner product with the query. We investigate the capacity of this architecture relative to sparse bag-of-words models and attentional neural networks. Using both theoretical and empirical analysis, we establish connections between the encoding dimension, the margin between gold and lower-ranked documents, and the document length, suggesting limitations in the capacity of fixed-length encodings to support precise retrieval of long documents. Building on these insights, we propose a simple neural model that combines the efficiency of dual encoders with some of the expressiveness of more costly attentional architectures, and explore sparse-dense hybrids to capitalize on the precision of sparse retrieval. These models outperform strong alternatives in large-scale retrieval.

We introduce an Edit-Based TransfOrmer with Repositioning (EDITOR), which makes sequence generation flexible by seamlessly allowing users to specify preferences in output lexical choice. Building on recent models for non-autoregressive sequence generation (Gu et al., 2019), EDITOR generates new sequences by iteratively editing hypotheses. It relies on a novel reposition operation designed to disentangle lexical choice from word positioning decisions, while enabling efficient oracles for imitation learning and parallel edits at decoding time. Empirically, EDITOR uses soft lexical constraints more effectively than the Levenshtein Transformer (Gu et al., 2019) while speeding up decoding dramatically compared to constrained beam search (Post and Vilar, 2018). EDITOR also achieves comparable or better translation quality with faster decoding speed than the Levenshtein Transformer on standard Romanian-English, English-German, and English-Japanese machine translation tasks.

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 present the Quantized Transformer (QT), an unsupervised system for extractive opinion summarization. QT is inspired by Vector- Quantized Variational Autoencoders, which we repurpose for popularity-driven summarization. It uses a clustering interpretation of the quantized space and a novel extraction algorithm to discover popular opinions among hundreds of reviews, a significant step towards opinion summarization of practical scope. In addition, QT enables controllable summarization without further training, by utilizing properties of the quantized space to extract aspect-specific summaries. We also make publicly available Space, a large-scale evaluation benchmark for opinion summarizers, comprising general and aspect-specific summaries for 50 hotels. Experiments demonstrate the promise of our approach, which is validated by human studies where judges showed clear preference for our method over competitive baselines.

Prior studies in multilingual language modeling (e.g., Cotterell et al., 2018; Mielke et al., 2019) disagree on whether or not inflectional morphology makes languages harder to model. We attempt to resolve the disagreement and extend those studies. We compile a larger corpus of 145 Bible translations in 92 languages and a larger number of typological features.1 We fill in missing typological data for several languages and consider corpus-based measures of morphological complexity in addition to expert-produced typological features. We find that several morphological measures are significantly associated with higher surprisal when LSTM models are trained with BPE-segmented data. We also investigate linguistically motivated subword segmentation strategies like Morfessor and Finite-State Transducers (FSTs) and find that these segmentation strategies yield better performance and reduce the impact of a language’s morphology on language modeling.

Although current CCG supertaggers achieve high accuracy on the standard WSJ test set, few systems make use of the categories’ internal structure that will drive the syntactic derivation during parsing. The tagset is traditionally truncated, discarding the many rare and complex category types in the long tail. However, supertags are themselves trees. Rather than give up on rare tags, we investigate constructive models that account for their internal structure, including novel methods for tree-structured prediction. Our best tagger is capable of recovering a sizeable fraction of the long-tail supertags and even generates CCG categories that have never been seen in training, while approximating the prior state of the art in overall tag accuracy with fewer parameters. We further investigate how well different approaches generalize to out-of-domain evaluation sets.

For natural language processing systems, two kinds of evidence support the use of text representations from neural language models “pretrained” on large unannotated corpora: performance on application-inspired benchmarks (Peters et al., 2018, inter alia), and the emergence of syntactic abstractions in those representations (Tenney et al., 2019, inter alia). On the other hand, the lack of grounded supervision calls into question how well these representations can ever capture meaning (Bender and Koller, 2020). We apply novel probes to recent language models— specifically focusing on predicate-argument structure as operationalized by semantic dependencies (Ivanova et al., 2012)—and find that, unlike syntax, semantics is not brought to the surface by today’s pretrained models. We then use convolutional graph encoders to explicitly incorporate semantic parses into task-specific finetuning, yielding benefits to natural language understanding (NLU) tasks in the GLUE benchmark. This approach demonstrates the potential for general-purpose (rather than task-specific) linguistic supervision, above and beyond conventional pretraining and finetuning. Several diagnostics help to localize the benefits of our approach.1