Machine Reading Comprehension: The Role of Contextualized Language Models and Beyond

Machine reading comprehension (MRC) aims to teach machines to read and comprehend human languages, which is a long-standing goal of natural language processing (NLP). With the burst of deep neural networks and the evolution of contextualized language models (CLMs), the research of MRC has experienced two significant breakthroughs. MRC and CLM, as a phenomenon, have a great impact on the NLP community. In this survey, we provide a comprehensive and comparative review on MRC covering overall research topics about 1) the origin and development of MRC and CLM, with a particular focus on the role of CLMs; 2) the impact of MRC and CLM to the NLP community; 3) the definition, datasets, and evaluation of MRC; 4) general MRC architecture and technical methods in the view of two-stage Encoder-Decoder solving architecture from the insights of the cognitive process of humans; 5) previous highlights, emerging topics, and our empirical analysis, among which we especially focus on what works in different periods of MRC researches. We propose a full-view categorization and new taxonomies on these topics. The primary views we have arrived at are that 1) MRC boosts the progress from language processing to understanding; 2) the rapid improvement of MRC systems greatly benefits from the development of CLMs; 3) the theme of MRC is gradually moving from shallow text matching to cognitive reasoning.

Bipartite Flat-Graph Network for Nested Named Entity Recognition

In this paper, we propose a novel bipartite flat-graph network (BiFlaG) for nested named entity recognition (NER), which contains two subgraph modules: a flat NER module for outermost entities and a graph module for all the entities located in inner layers. Bidirectional LSTM (BiLSTM) and graph convolutional network (GCN) are adopted to jointly learn flat entities and their inner dependencies. Different from previous models, which only consider the unidirectional delivery of information from innermost layers to outer ones (or outside-to-inside), our model effectively captures the bidirectional interaction between them. We first use the entities recognized by the flat NER module to construct an entity graph, which is fed to the next graph module. The richer representation learned from graph module carries the dependencies of inner entities and can be exploited to improve outermost entity predictions. Experimental results on three standard nested NER datasets demonstrate that our BiFlaG outperforms previous state-of-the-art models.

Capsule-Transformer for Neural Machine Translation

Transformer hugely benefits from its key design of the multi-head self-attention network (SAN), which extracts information from various perspectives through transforming the given input into different subspaces. However, its simple linear transformation aggregation strategy may still potentially fail to fully capture deeper contextualized information. In this paper, we thus propose the capsule-Transformer, which extends the linear transformation into a more general capsule routing algorithm by taking SAN as a special case of capsule network. So that the resulted capsule-Transformer is capable of obtaining a better attention distribution representation of the input sequence via information aggregation among different heads and words. Specifically, we see groups of attention weights in SAN as low layer capsules. By applying the iterative capsule routing algorithm they can be further aggregated into high layer capsules which contain deeper contextualized information. Experimental results on the widely-used machine translation datasets show our proposed capsule-Transformer outperforms strong Transformer baseline significantly.

Syntax-aware Data Augmentation for Neural Machine Translation

Data augmentation is an effective performance enhancement in neural machine translation (NMT) by generating additional bilingual data. In this paper, we propose a novel data augmentation enhancement strategy for neural machine translation. Different from existing data augmentation methods which simply choose words with the same probability across different sentences for modification, we set sentence-specific probability for word selection by considering their roles in sentence. We use dependency parse tree of input sentence as an effective clue to determine selecting probability for every words in each sentence. Our proposed method is evaluated on WMT14 English-to-German dataset and IWSLT14 German-to-English dataset. The result of extensive experiments show our proposed syntax-aware data augmentation method may effectively boost existing sentence-independent methods for significant translation performance improvement.

Multi-choice Dialogue-Based Reading Comprehension with Knowledge and Key Turns

Multi-choice machine reading comprehension (MRC) requires models to choose the correct answer from candidate options given a passage and a question. Our research focuses dialogue-based MRC, where the passages are multi-turn dialogues. It suffers from two challenges, the answer selection decision is made without support of latently helpful commonsense, and the multi-turn context may hide considerable irrelevant information. This work thus makes the first attempt to tackle those two challenges by extracting substantially important turns and utilizing external knowledge to enhance the representation of context. In this paper, the relevance of each turn to the question are calculated to choose key turns. Besides, terms related to the context and the question in a knowledge graph are extracted as external knowledge. The original context, question and external knowledge are encoded with the pre-trained language model, then the language representation and key turns are combined together with a will-designed mechanism to predict the answer. Experimental results on a DREAM dataset show that our proposed model achieves great improvements on baselines.

Learning Better Universal Representations from Pre-trained Contextualized Language Models

Pre-trained contextualized language models such as BERT have shown great effectiveness in a wide range of downstream natural language processing (NLP) tasks. However, the effective representations offered by the models target at each token inside a sequence rather than each sequence and the fine-tuning step involves the input of both sequences at one time, leading to unsatisfying representation of each individual sequence. Besides, as sentence-level representations taken as the full training context in these models, there comes inferior performance on lower-level linguistic units (phrases and words). In this work, we present a novel framework on BERT that is capable of generating universal, fixed-size representations for input sequences of any lengths, i.e., words, phrases, and sentences, using a large scale of natural language inference and paraphrase data with multiple training objectives. Our proposed framework adopts the Siamese network, learning sentence-level representations from natural language inference dataset and phrase and word-level representations from paraphrasing dataset, respectively. We evaluate our model across different granularity of text similarity tasks, including STS tasks, SemEval2013 Task 5(a) and some commonly used word similarity tasks, where our model substantially outperforms other representation models on sentence-level datasets and achieves significant improvements in word-level and phrase-level representation.

Semantics-Aware Inferential Network for Natural Language Understanding

For natural language understanding tasks, either machine reading comprehension or natural language inference, both semantics-aware and inference are favorable features of the concerned modeling for better understanding performance. Thus we propose a Semantics-Aware Inferential Network (SAIN) to meet such a motivation. Taking explicit contextualized semantics as a complementary input, the inferential module of SAIN enables a series of reasoning steps over semantic clues through an attention mechanism. By stringing these steps, the inferential network effectively learns to perform iterative reasoning which incorporates both explicit semantics and contextualized representations. In terms of well pre-trained language models as front-end encoder, our model achieves significant improvement on 11 tasks including machine reading comprehension and natural language inference.

Unsupervised Neural Machine Translation with Indirect Supervision

Neural machine translation~(NMT) is ineffective for zero-resource languages. Recent works exploring the possibility of unsupervised neural machine translation (UNMT) with only monolingual data can achieve promising results. However, there are still big gaps between UNMT and NMT with parallel supervision. In this work, we introduce a multilingual unsupervised NMT (\method) framework to leverage weakly supervised signals from high-resource language pairs to zero-resource translation directions. More specifically, for unsupervised language pairs \texttt{En-De}, we can make full use of the information from parallel dataset \texttt{En-Fr} to jointly train the unsupervised translation directions all in one model. \method is based on multilingual models which require no changes to the standard unsupervised NMT. Empirical results demonstrate that \method significantly improves the translation quality by more than 3 BLEU score on six benchmark unsupervised translation directions.

Reference Language based Unsupervised Neural Machine Translation

Exploiting common language as an auxiliary for better translation has a long tradition in machine translation, which lets supervised learning based machine translation enjoy the enhancement delivered by the well-used pivot language, in case that the prerequisite of parallel corpus from source language to target language cannot be fully satisfied. The rising of unsupervised neural machine translation (UNMT) seems completely relieving the parallel corpus curse, though still subject to unsatisfactory performance so far due to vague clues available used for its core back-translation training. Further enriching the idea of pivot translation by freeing the use of parallel corpus other than its specified source and target, we propose a new reference language based UNMT framework, in which the reference language only shares parallel corpus with the source, indicating clear enough signal to help the reconstruction training of UNMT through a proposed reference agreement mechanism. Experimental results show that our methods improve the quality of UNMT over that of a strong baseline in terms of only one auxiliary language, demonstrating the usefulness of the proposed reference language based UNMT with a good start.

Dual Multi-head Co-attention for Multi-choice Reading Comprehension

Multi-choice Machine Reading Comprehension (MRC) requires model to decide the correct answer from a set of answer options when given a passage and a question. Thus in addition to a powerful pre-trained Language Model as encoder, multi-choice MRC especially relies on a matching network design which is supposed to effectively capture the relationship among the triplet of passage, question and answers. While the latest pre-trained Language Models have shown powerful enough even without the support from a matching network, and the latest matching network has been complicated enough, we thus propose a novel going-back-to-the-basic solution which straightforwardly models the MRC relationship as attention mechanism inside network. The proposed DUal Multi-head Co-Attention (DUMA) has been shown simple but effective and is capable of generally promoting pre-trained Language Models. Our proposed method is evaluated on two benchmark multi-choice MRC tasks, DREAM and RACE, showing that in terms of strong Language Models, DUMA may still boost the model to reach new state-of-the-art performance.