ACL 2020 论文列表

Abstract Meaning Representations (AMRs) capture sentence-level semantics structural representations to broad-coverage natural sentences. We investigate parsing AMR with explicit dependency structures and interpretable latent structures. We generate the latent soft structure without additional annotations, and fuse both dependency and latent structure via an extended graph neural networks. The fused structural information helps our experiments results to achieve the best reported results on both AMR 2.0 (77.5% Smatch F1 on LDC2017T10) and AMR 1.0 ((71.8% Smatch F1 on LDC2014T12).

Moderation is crucial to promoting healthy online discussions. Although several ‘toxicity’ detection datasets and models have been published, most of them ignore the context of the posts, implicitly assuming that comments may be judged independently. We investigate this assumption by focusing on two questions: (a) does context affect the human judgement, and (b) does conditioning on context improve performance of toxicity detection systems? We experiment with Wikipedia conversations, limiting the notion of context to the previous post in the thread and the discussion title. We find that context can both amplify or mitigate the perceived toxicity of posts. Moreover, a small but significant subset of manually labeled posts (5% in one of our experiments) end up having the opposite toxicity labels if the annotators are not provided with context. Surprisingly, we also find no evidence that context actually improves the performance of toxicity classifiers, having tried a range of classifiers and mechanisms to make them context aware. This points to the need for larger datasets of comments annotated in context. We make our code and data publicly available.

The key to effortless end-user programming is natural language. We examine how to teach intelligent systems new functions, expressed in natural language. As a first step, we collected 3168 samples of teaching efforts in plain English. Then we built fuSE, a novel system that translates English function descriptions into code. Our approach is three-tiered and each task is evaluated separately. We first classify whether an intent to teach new functionality is present in the utterance (accuracy: 97.7% using BERT). Then we analyze the linguistic structure and construct a semantic model (accuracy: 97.6% using a BiLSTM). Finally, we synthesize the signature of the method, map the intermediate steps (instructions in the method body) to API calls and inject control structures (F1: 67.0% with information retrieval and knowledge-based methods). In an end-to-end evaluation on an unseen dataset fuSE synthesized 84.6% of the method signatures and 79.2% of the API calls correctly.

The rise of online communication platforms has been accompanied by some undesirable effects, such as the proliferation of aggressive and abusive behaviour online. Aiming to tackle this problem, the natural language processing (NLP) community has experimented with a range of techniques for abuse detection. While achieving substantial success, these methods have so far only focused on modelling the linguistic properties of the comments and the online communities of users, disregarding the emotional state of the users and how this might affect their language. The latter is, however, inextricably linked to abusive behaviour. In this paper, we present the first joint model of emotion and abusive language detection, experimenting in a multi-task learning framework that allows one task to inform the other. Our results demonstrate that incorporating affective features leads to significant improvements in abuse detection performance across datasets.

In this paper, we study the problem of identifying the principals and accessories from the fact description with multiple defendants in a criminal case. We treat the fact descriptions as narrative texts and the defendants as roles over the narrative story. We propose to model the defendants with behavioral semantic information and statistical characteristics, then learning the importances of defendants within a learning-to-rank framework. Experimental results on a real-world dataset demonstrate the behavior analysis can effectively model the defendants’ impacts in a complex case.

With the explosion of news information, personalized news recommendation has become very important for users to quickly find their interested contents. Most existing methods usually learn the representations of users and news from news contents for recommendation. However, they seldom consider high-order connectivity underlying the user-news interactions. Moreover, existing methods failed to disentangle a user’s latent preference factors which cause her clicks on different news. In this paper, we model the user-news interactions as a bipartite graph and propose a novel Graph Neural News Recommendation model with Unsupervised Preference Disentanglement, named GNUD. Our model can encode high-order relationships into user and news representations by information propagation along the graph. Furthermore, the learned representations are disentangled with latent preference factors by a neighborhood routing algorithm, which can enhance expressiveness and interpretability. A preference regularizer is also designed to force each disentangled subspace to independently reflect an isolated preference, improving the quality of the disentangled representations. Experimental results on real-world news datasets demonstrate that our proposed model can effectively improve the performance of news recommendation and outperform state-of-the-art news recommendation methods.

This paper investigates how to effectively incorporate a pre-trained masked language model (MLM), such as BERT, into an encoder-decoder (EncDec) model for grammatical error correction (GEC). The answer to this question is not as straightforward as one might expect because the previous common methods for incorporating a MLM into an EncDec model have potential drawbacks when applied to GEC. For example, the distribution of the inputs to a GEC model can be considerably different (erroneous, clumsy, etc.) from that of the corpora used for pre-training MLMs; however, this issue is not addressed in the previous methods. Our experiments show that our proposed method, where we first fine-tune a MLM with a given GEC corpus and then use the output of the fine-tuned MLM as additional features in the GEC model, maximizes the benefit of the MLM. The best-performing model achieves state-of-the-art performances on the BEA-2019 and CoNLL-2014 benchmarks. Our code is publicly available at: https://github.com/kanekomasahiro/bert-gec.

Existing approaches to active learning maximize the system performance by sampling unlabeled instances for annotation that yield the most efficient training. However, when active learning is integrated with an end-user application, this can lead to frustration for participating users, as they spend time labeling instances that they would not otherwise be interested in reading. In this paper, we propose a new active learning approach that jointly optimizes the seemingly counteracting objectives of the active learning system (training efficiently) and the user (receiving useful instances). We study our approach in an educational application, which particularly benefits from this technique as the system needs to rapidly learn to predict the appropriateness of an exercise to a particular user, while the users should receive only exercises that match their skills. We evaluate multiple learning strategies and user types with data from real users and find that our joint approach better satisfies both objectives when alternative methods lead to many unsuitable exercises for end users.

This paper studies strategies to model word formation in NMT using rich linguistic information, namely a word segmentation approach that goes beyond splitting into substrings by considering fusional morphology. Our linguistically sound segmentation is combined with a method for target-side inflection to accommodate modeling word formation. The best system variants employ source-side morphological analysis and model complex target-side words, improving over a standard system.

Multi-task Learning methods have achieved great progress in text classification. However, existing methods assume that multi-task text classification problems are convex multiobjective optimization problems, which is unrealistic in real-world applications. To address this issue, this paper presents a novel Tchebycheff procedure to optimize the multi-task classification problems without convex assumption. The extensive experiments back up our theoretical analysis and validate the superiority of our proposals.

Recent studies on interpretability of attention distributions have led to notions of faithful and plausible explanations for a model’s predictions. Attention distributions can be considered a faithful explanation if a higher attention weight implies a greater impact on the model’s prediction. They can be considered a plausible explanation if they provide a human-understandable justification for the model’s predictions. In this work, we first explain why current attention mechanisms in LSTM based encoders can neither provide a faithful nor a plausible explanation of the model’s predictions. We observe that in LSTM based encoders the hidden representations at different time-steps are very similar to each other (high conicity) and attention weights in these situations do not carry much meaning because even a random permutation of the attention weights does not affect the model’s predictions. Based on experiments on a wide variety of tasks and datasets, we observe attention distributions often attribute the model’s predictions to unimportant words such as punctuation and fail to offer a plausible explanation for the predictions. To make attention mechanisms more faithful and plausible, we propose a modified LSTM cell with a diversity-driven training objective that ensures that the hidden representations learned at different time steps are diverse. We show that the resulting attention distributions offer more transparency as they (i) provide a more precise importance ranking of the hidden states (ii) are better indicative of words important for the model’s predictions (iii) correlate better with gradient-based attribution methods. Human evaluations indicate that the attention distributions learned by our model offer a plausible explanation of the model’s predictions. Our code has been made publicly available at https://github.com/akashkm99/Interpretable-Attention

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

In the Transformer model, “self-attention” combines information from attended embeddings into the representation of the focal embedding in the next layer. Thus, across layers of the Transformer, information originating from different tokens gets increasingly mixed. This makes attention weights unreliable as explanations probes. In this paper, we consider the problem of quantifying this flow of information through self-attention. We propose two methods for approximating the attention to input tokens given attention weights, attention rollout and attention flow, as post hoc methods when we use attention weights as the relative relevance of the input tokens. We show that these methods give complementary views on the flow of information, and compared to raw attention, both yield higher correlations with importance scores of input tokens obtained using an ablation method and input gradients.

Language models keep track of complex information about the preceding context – including, e.g., syntactic relations in a sentence. We investigate whether they also capture information beneficial for resolving pronominal anaphora in English. We analyze two state of the art models with LSTM and Transformer architectures, via probe tasks and analysis on a coreference annotated corpus. The Transformer outperforms the LSTM in all analyses. Our results suggest that language models are more successful at learning grammatical constraints than they are at learning truly referential information, in the sense of capturing the fact that we use language to refer to entities in the world. However, we find traces of the latter aspect, too.

By introducing a small set of additional parameters, a probe learns to solve specific linguistic tasks (e.g., dependency parsing) in a supervised manner using feature representations (e.g., contextualized embeddings). The effectiveness of such probing tasks is taken as evidence that the pre-trained model encodes linguistic knowledge. However, this approach of evaluating a language model is undermined by the uncertainty of the amount of knowledge that is learned by the probe itself. Complementary to those works, we propose a parameter-free probing technique for analyzing pre-trained language models (e.g., BERT). Our method does not require direct supervision from the probing tasks, nor do we introduce additional parameters to the probing process. Our experiments on BERT show that syntactic trees recovered from BERT using our method are significantly better than linguistically-uninformed baselines. We further feed the empirically induced dependency structures into a downstream sentiment classification task and find its improvement compatible with or even superior to a human-designed dependency schema.

To increase trust in artificial intelligence systems, a promising research direction consists of designing neural models capable of generating natural language explanations for their predictions. In this work, we show that such models are nonetheless prone to generating mutually inconsistent explanations, such as ”Because there is a dog in the image.” and ”Because there is no dog in the [same] image.”, exposing flaws in either the decision-making process of the model or in the generation of the explanations. We introduce a simple yet effective adversarial framework for sanity checking models against the generation of inconsistent natural language explanations. Moreover, as part of the framework, we address the problem of adversarial attacks with full target sequences, a scenario that was not previously addressed in sequence-to-sequence attacks. Finally, we apply our framework on a state-of-the-art neural natural language inference model that provides natural language explanations for its predictions. Our framework shows that this model is capable of generating a significant number of inconsistent explanations.

Given the fast development of analysis techniques for NLP and speech processing systems, few systematic studies have been conducted to compare the strengths and weaknesses of each method. As a step in this direction we study the case of representations of phonology in neural network models of spoken language. We use two commonly applied analytical techniques, diagnostic classifiers and representational similarity analysis, to quantify to what extent neural activation patterns encode phonemes and phoneme sequences. We manipulate two factors that can affect the outcome of analysis. First, we investigate the role of learning by comparing neural activations extracted from trained versus randomly-initialized models. Second, we examine the temporal scope of the activations by probing both local activations corresponding to a few milliseconds of the speech signal, and global activations pooled over the whole utterance. We conclude that reporting analysis results with randomly initialized models is crucial, and that global-scope methods tend to yield more consistent and interpretable results and we recommend their use as a complement to local-scope diagnostic methods.

With the recent proliferation of the use of text classifications, researchers have found that there are certain unintended biases in text classification datasets. For example, texts containing some demographic identity-terms (e.g., “gay”, “black”) are more likely to be abusive in existing abusive language detection datasets. As a result, models trained with these datasets may consider sentences like “She makes me happy to be gay” as abusive simply because of the word “gay.” In this paper, we formalize the unintended biases in text classification datasets as a kind of selection bias from the non-discrimination distribution to the discrimination distribution. Based on this formalization, we further propose a model-agnostic debiasing training framework by recovering the non-discrimination distribution using instance weighting, which does not require any extra resources or annotations apart from a pre-defined set of demographic identity-terms. Experiments demonstrate that our method can effectively alleviate the impacts of the unintended biases without significantly hurting models’ generalization ability.

Recent work has shown that neural rerankers can improve results for dependency parsing over the top k trees produced by a base parser. However, all neural rerankers so far have been evaluated on English and Chinese only, both languages with a configurational word order and poor morphology. In the paper, we re-assess the potential of successful neural reranking models from the literature on English and on two morphologically rich(er) languages, German and Czech. In addition, we introduce a new variation of a discriminative reranker based on graph convolutional networks (GCNs). We show that the GCN not only outperforms previous models on English but is the only model that is able to improve results over the baselines on German and Czech. We explain the differences in reranking performance based on an analysis of a) the gold tree ratio and b) the variety in the k-best lists.