COLING 2020 Papers

Pre-training sentence encoders is effective in many natural language processing tasks including machine translation (MT) quality estimation (QE), due partly to the scarcity of annotated QE data required for supervised learning. In this paper, we investigate the use of an intermediate self-supervised learning task for sentence encoder aiming at improving QE performances at the sentence and word levels. Our approach is motivated by a problem inherent to QE: mistakes in translation caused by wrongly inserted and deleted tokens. We modify the translation language model (TLM) training objective of the cross-lingual language model (XLM) to orientate the pre-trained model towards the target task. The proposed method does not rely on annotated data and is complementary to QE methods involving pre-trained sentence encoders and domain adaptation. Experiments on English-to-German and English-to-Russian translation directions show that intermediate learning improves over domain adaptated models. Additionally, our method reaches results in par with state-of-the-art QE models without requiring the combination of several approaches and outperforms similar methods based on pre-trained sentence encoders.

We tackle implicit discourse relation recognition. Both self-attention and interactive-attention mechanisms have been applied for attention-aware representation learning, which improves the current discourse analysis models. To take advantages of the two attention mechanisms simultaneously, we develop a propagative attention learning model using a cross-coupled two-channel network. We experiment on Penn Discourse Treebank. The test results demonstrate that our model yields substantial improvements over the baselines (BiLSTM and BERT).

Text input technologies for low-resource languages support literacy, content authoring, and language learning. However, tasks such as word completion pose a challenge for morphologically complex languages thanks to the combinatorial explosion of possible words. We have developed a method for morphologically-aware text input in Kunwinjku, a polysynthetic language of northern Australia. We modify an existing finite state recognizer to map input morph prefixes to morph completions, respecting the morphosyntax and morphophonology of the language. We demonstrate the portability of the method by applying it to Turkish. We show that the space of proximal morph completions is many orders of magnitude smaller than the space of full word completions for Kunwinjku. We provide a visualization of the morph completion space to enable the text completion parameters to be fine-tuned. Finally, we report on a web services deployment, along with a web interface which helps users enter morphologically complex words and which retrieves corresponding entries from the lexicon.

Coping with ambiguous questions has been a perennial problem in real-world dialogue systems. Although clarification by asking questions is a common form of human interaction, it is hard to define appropriate questions to elicit more specific intents from a user. In this work, we propose a reinforcement model to clarify ambiguous questions by suggesting refinements of the original query. We first formulate a collection partitioning problem to select a set of labels enabling us to distinguish potential unambiguous intents. We list the chosen labels as intent phrases to the user for further confirmation. The selected label along with the original user query then serves as a refined query, for which a suitable response can more easily be identified. The model is trained using reinforcement learning with a deep policy network. We evaluate our model based on real-world user clicks and demonstrate significant improvements across several different experiments.

Image paragraph captioning (IPC) aims to generate a fine-grained paragraph to describe the visual content of an image. Significant progress has been made by deep neural networks, in which the attention mechanism plays an essential role. However, conventional attention mechanisms tend to ignore the past alignment information, which often results in problems of repetitive captioning and incomplete captioning. In this paper, we propose an Interactive key-value Memory- augmented Attention model for image Paragraph captioning (IMAP) to keep track of the attention history (salient objects coverage information) along with the update-chain of the decoder state and therefore avoid generating repetitive or incomplete image descriptions. In addition, we employ an adaptive attention mechanism to realize adaptive alignment from image regions to caption words, where an image region can be mapped to an arbitrary number of caption words while a caption word can also attend to an arbitrary number of image regions. Extensive experiments on a benchmark dataset (i.e., Stanford) demonstrate the effectiveness of our IMAP model.

Conversational systems are of primary interest in the AI community. Organizations are increasingly using chatbot to provide round-the-clock support and to increase customer engagement. Many commercial bot building frameworks follow a standard approach that requires one to build and train an intent model to recognize user input. These frameworks require a collection of user utterances and corresponding intent to train an intent model. Collecting a substantial coverage of training data is a bottleneck in the bot building process. In cases where past conversation data is available, the cost of labeling hundreds of utterances with intent labels is time-consuming and laborious. In this paper, we present an intent discovery framework that can mine a vast amount of conversational logs and to generate labeled data sets for training intent models. We have introduced an extension to the DBSCAN algorithm and presented a density-based clustering algorithm ITER-DBSCAN for unbalanced data clustering. Empirical evaluation on one conversation dataset, six different intent dataset, and one short text clustering dataset show the effectiveness of our hypothesis.

Dialogue Act Recognition (DAR) is a challenging problem in Natural Language Understanding, which aims to attach Dialogue Act (DA) labels to each utterance in a conversation. However, previous studies cannot fully recognize the specific expressions given by users due to the informality and diversity of natural language expressions. To solve this problem, we propose a Heterogeneous User History (HUH) graph convolution network, which utilizes the user’s historical answers grouped by DA labels as additional clues to recognize the DA label of utterances. To handle the noise caused by introducing the user’s historical answers, we design sets of denoising mechanisms, including a History Selection process, a Similarity Re-weighting process, and an Edge Re-weighting process. We evaluate the proposed method on two benchmark datasets MSDialog and MRDA. The experimental results verify the effectiveness of integrating user’s historical answers, and show that our proposed model outperforms the state-of-the-art methods.

Script learning aims to predict the subsequent event according to the existing event chain. Recent studies focus on event co-occurrence to solve this problem. However, few studies integrate external event knowledge to solve this problem. With our observations, external event knowledge can provide additional knowledge like temporal or causal knowledge for understanding event chain better and predicting the right subsequent event. In this work, we integrate event knowledge from ASER (Activities, States, Events and their Relations) knowledge base to help predict the next event. We propose a new approach consisting of knowledge retrieval stage and knowledge integration stage. In the knowledge retrieval stage, we select relevant external event knowledge from ASER. In the knowledge integration stage, we propose three methods to integrate external knowledge into our model and infer final answers. Experiments on the widely-used Multi- Choice Narrative Cloze (MCNC) task show our approach achieves state-of-the-art performance compared to other methods.

This paper extends existing work on terminology integration into Neural Machine Translation, a common industrial practice to dynamically adapt translation to a specific domain. Our method, based on the use of placeholders complemented with morphosyntactic annotation, efficiently taps into the ability of the neural network to deal with symbolic knowledge to surpass the surface generalization shown by alternative techniques. We compare our approach to state-of-the-art systems and benchmark them through a well-defined evaluation framework, focusing on actual application of terminology and not just on the overall performance. Results indicate the suitability of our method in the use-case where terminology is used in a system trained on generic data only.

Knowledge bases (KBs) are essential for many downstream NLP tasks, yet their prime shortcoming is that they are often incomplete. State-of-the-art frameworks for KB completion often lack sufficient accuracy to work fully automated without human supervision. As a remedy, we propose : a novel interactive framework for KB completion from text based on a question answering pipeline. Our framework is tailored to the specific needs of a human-in-the-loop paradigm: (i) We generate facts that are aligned with text snippets and are thus immediately verifiable by humans. (ii) Our system is designed such that it continuously learns during the KB completion task and, therefore, significantly improves its performance upon initial zero- and few-shot relations over time. (iii) We only trigger human interactions when there is enough information for a correct prediction. Therefore, we train our system with negative examples and a fold-option if there is no answer. Our framework yields a favorable performance: it achieves a hit@1 ratio of 29.7% for initially unseen relations, upon which it gradually improves to 46.2%.

Non-autoregressive models generate target words in a parallel way, which achieve a faster decoding speed but at the sacrifice of translation accuracy. To remedy a flawed translation by non-autoregressive models, a promising approach is to train a conditional masked translation model (CMTM), and refine the generated results within several iterations. Unfortunately, such approach hardly considers the sequential dependency among target words, which inevitably results in a translation degradation. Hence, instead of solely training a Transformer-based CMTM, we propose a Self-Review Mechanism to infuse sequential information into it. Concretely, we insert a left-to-right mask to the same decoder of CMTM, and then induce it to autoregressively review whether each generated word from CMTM is supposed to be replaced or kept. The experimental results (WMT14 En ↔ De and WMT16 En ↔ Ro) demonstrate that our model uses dramatically less training computations than the typical CMTM, as well as outperforms several state-of-the-art non-autoregressive models by over 1 BLEU. Through knowledge distillation, our model even surpasses a typical left-to-right Transformer model, while significantly speeding up decoding.

This work proposes a new method for manual evaluation of Machine Translation (MT) output based on marking actual issues in the translated text. The novelty is that the evaluators are not assigning any scores, nor classifying errors, but marking all problematic parts (words, phrases, sentences) of the translation. The main advantage of this method is that the resulting annotations do not only provide overall scores by counting words with assigned tags, but can be further used for analysis of errors and challenging linguistic phenomena, as well as inter-annotator disagreements. Detailed analysis and understanding of actual problems are not enabled by typical manual evaluations where the annotators are asked to assign overall scores or to rank two or more translations. The proposed method is very general: it can be applied on any genre/domain and language pair, and it can be guided by various types of quality criteria. Also, it is not restricted to MT output, but can be used for other types of generated text.

We investigate the impact of political ideology biases in training data. Through a set of comparison studies, we examine the propagation of biases in several widely-used NLP models and its effect on the overall retrieval accuracy. Our work highlights the susceptibility of large, complex models to propagating the biases from human-selected input, which may lead to a deterioration of retrieval accuracy, and the importance of controlling for these biases. Finally, as a way to mitigate the bias, we propose to learn a text representation that is invariant to political ideology while still judging topic relevance.

Although the Indonesian language is spoken by almost 200 million people and the 10th most spoken language in the world, it is under-represented in NLP research. Previous work on Indonesian has been hampered by a lack of annotated datasets, a sparsity of language resources, and a lack of resource standardization. In this work, we release the IndoLEM dataset comprising seven tasks for the Indonesian language, spanning morpho-syntax, semantics, and discourse. We additionally release IndoBERT, a new pre-trained language model for Indonesian, and evaluate it over IndoLEM, in addition to benchmarking it against existing resources. Our experiments show that IndoBERT achieves state-of-the-art performance over most of the tasks in IndoLEM.

Pretrained language models, neural models pretrained on massive amounts of data, have established the state of the art in a range of NLP tasks. They are based on a modern machine-learning technique, the Transformer which relates all items simultaneously to capture semantic relations in sequences. However, it differs from what humans do. Humans read sentences one-by-one, incrementally. Can neural models benefit by interpreting texts incrementally as humans do? We investigate this question in coherence modeling. We propose a coherence model which interprets sentences incrementally to capture lexical relations between them. We compare the state of the art in each task, simple neural models relying on a pretrained language model, and our model in two downstream tasks. Our findings suggest that interpreting texts incrementally as humans could be useful to design more advanced models.

Self-Attention Networks (SANs) are an integral part of successful neural architectures such as Transformer (Vaswani et al., 2017), and thus of pretrained language models such as BERT (Devlin et al., 2019) or GPT-3 (Brown et al., 2020). Training SANs on a task or pretraining them on language modeling requires large amounts of data and compute resources. We are searching for modifications to SANs that enable faster learning, i.e., higher accuracies after fewer update steps. We investigate three modifications to SANs: direct position interactions, learnable temperature, and convoluted attention. When evaluating them on part-of-speech tagging, we find that direct position interactions are an alternative to position embeddings, and convoluted attention has the potential to speed up the learning process.

Machine reading comprehension (MRC) is one of the most critical yet challenging tasks in natural language understanding(NLU), where both syntax and semantics information of text are essential components for text understanding. It is surprising that jointly considering syntax and semantics in neural networks was never formally reported in literature. This paper makes the first attempt by proposing a novel Syntax and Frame Semantics model for Machine Reading Comprehension (SS-MRC), which takes full advantage of syntax and frame semantics to get richer text representation. Our extensive experimental results demonstrate that SS-MRC performs better than ten state-of-the-art technologies on machine reading comprehension task.

Neural Machine Translation often suffers from an under-translation problem due to its limited modeling of output sequence lengths. In this work, we propose a novel approach to training a Transformer model using length constraints based on length-aware positional encoding (PE). Since length constraints with exact target sentence lengths degrade translation performance, we add random noise within a certain window size to the length constraints in the PE during the training. In the inference step, we predict the output lengths using input sequences and a BERT-based length prediction model. Experimental results in an ASPEC English-to-Japanese translation showed the proposed method produced translations with lengths close to the reference ones and outperformed a vanilla Transformer (especially in short sentences) by 3.22 points in BLEU. The average translation results using our length prediction model were also better than another baseline method using input lengths for the length constraints. The proposed noise injection improved robustness for length prediction errors, especially within the window size.

Mongolian morphological segmentation is regarded as a crucial preprocessing step in many Mongolian related NLP applications and has received extensive attention. Recently, end-to-end segmentation approaches with long short-term memory networks (LSTM) have achieved excellent results. However, the inner-word features among characters in the word and the out-word features from context are not well utilized in the segmentation process. In this paper, we propose a neural network incorporating inner-word and out-word features for Mongolian morphological segmentation. The network consists of two encoders and one decoder. The inner-word encoder uses the self-attention mechanisms to capture the inner-word features of the target word. The out-word encoder employs a two layers BiLSTM network to extract out-word features in the sentence. Then, the decoder adopts a multi-head double attention layer to fuse the inner-word features and out-word features and produces the segmentation result. The evaluation experiment compares the proposed network with the baselines and explores the effectiveness of the sub-modules.