Table-to-Text Generation with Effective Hierarchical Encoder on Three Dimensions (Row, Column and Time)

Although Seq2Seq models for table-to-text generation have achieved remarkable progress, modeling table representation in one dimension is inadequate. This is because (1) the table consists of multiple rows and columns, which means that encoding a table should not depend only on one dimensional sequence or set of records and (2) most of the tables are time series data (e.g. NBA game data, stock market data), which means that the description of the current table may be affected by its historical data. To address aforementioned problems, not only do we model each table cell considering other records in the same row, we also enrich table's representation by modeling each table cell in context of other cells in the same column or with historical (time dimension) data respectively. In addition, we develop a table cell fusion gate to combine representations from row, column and time dimension into one dense vector according to the saliency of each dimension's representation. We evaluated our methods on ROTOWIRE, a benchmark dataset of NBA basketball games. Both automatic and human evaluation results demonstrate the effectiveness of our model with improvement of 2.66 in BLEU over the strong baseline and outperformance of state-of-the-art model.

A Stack-Propagation Framework with Token-Level Intent Detection for Spoken Language Understanding

Intent detection and slot filling are two main tasks for building a spoken language understanding (SLU) system. The two tasks are closely tied and the slots often highly depend on the intent. In this paper, we propose a novel framework for SLU to better incorporate the intent information, which further guides the slot filling. In our framework, we adopt a joint model with Stack-Propagation which can directly use the intent information as input for slot filling, thus to capture the intent semantic knowledge. In addition, to further alleviate the error propagation, we perform the token-level intent detection for the Stack-Propagation framework. Experiments on two publicly datasets show that our model achieves the state-of-the-art performance and outperforms other previous methods by a large margin. Finally, we use the Bidirectional Encoder Representation from Transformer (BERT) model in our framework, which further boost our performance in SLU task.

Cross-Lingual Machine Reading Comprehension

Though the community has made great progress on Machine Reading Comprehension (MRC) task, most of the previous works are solving English-based MRC problems, and there are few efforts on other languages mainly due to the lack of large-scale training data. In this paper, we propose Cross-Lingual Machine Reading Comprehension (CLMRC) task for the languages other than English. Firstly, we present several back-translation approaches for CLMRC task, which is straightforward to adopt. However, to accurately align the answer into another language is difficult and could introduce additional noise. In this context, we propose a novel model called Dual BERT, which takes advantage of the large-scale training data provided by rich-resource language (such as English) and learn the semantic relations between the passage and question in a bilingual context, and then utilize the learned knowledge to improve reading comprehension performance of low-resource language. We conduct experiments on two Chinese machine reading comprehension datasets CMRC 2018 and DRCD. The results show consistent and significant improvements over various state-of-the-art systems by a large margin, which demonstrate the potentials in CLMRC task. Resources available: https://github.com/ymcui/Cross-Lingual-MRC

Multi-Task Self-Supervised Learning for Disfluency Detection

Most existing approaches to disfluency detection heavily rely on human-annotated data, which is expensive to obtain in practice. To tackle the training data bottleneck, we investigate methods for combining multiple self-supervised tasks-i.e., supervised tasks where data can be collected without manual labeling. First, we construct large-scale pseudo training data by randomly adding or deleting words from unlabeled news data, and propose two self-supervised pre-training tasks: (i) tagging task to detect the added noisy words. (ii) sentence classification to distinguish original sentences from grammatically-incorrect sentences. We then combine these two tasks to jointly train a network. The pre-trained network is then fine-tuned using human-annotated disfluency detection training data. Experimental results on the commonly used English Switchboard test set show that our approach can achieve competitive performance compared to the previous systems (trained using the full dataset) by using less than 1% (1000 sentences) of the training data. Our method trained on the full dataset significantly outperforms previous methods, reducing the error by 21% on English Switchboard.

ELG: An Event Logic Graph

The evolution and development of events have their own basic principles, which make events happen sequentially. Therefore, the discovery of such evolutionary patterns among events are of great value for event prediction, decision-making and scenario design of dialog systems. However, conventional knowledge graph mainly focuses on the entities and their relations, which neglects the real world events. In this paper, we present a novel type of knowledge base - Event Logic Graph (ELG), which can reveal evolutionary patterns and development logics of real world events. Specifically, ELG is a directed cyclic graph, whose nodes are events, and edges stand for the sequential, causal, conditional or hypernym-hyponym (is-a) relations between events. We constructed two domain ELG: financial domain ELG, which consists of more than 1.5 million of event nodes and more than 1.8 million of directed edges, and travel domain ELG, which consists of about 30 thousand of event nodes and more than 234 thousand of directed edges. Experimental results show that ELG is effective for the task of script event prediction.

Constructing Information-Lossless Biological Knowledge Graphs from Conditional Statements

Conditions are essential in the statements of biological literature. Without the conditions (e.g., environment, equipment) that were precisely specified, the facts (e.g., observations) in the statements may no longer be valid. One biological statement has one or multiple fact(s) and/or condition(s). Their subject and object can be either a concept or a concept's attribute. Existing information extraction methods do not consider the role of condition in the biological statement nor the role of attribute in the subject/object. In this work, we design a new tag schema and propose a deep sequence tagging framework to structure conditional statement into fact and condition tuples from biological text. Experiments demonstrate that our method yields a information-lossless structure of the literature.

Few-Shot Sequence Labeling with Label Dependency Transfer

Few-shot sequence labeling faces a unique challenge compared with the other fewshot classification problems, owing to the necessity for modeling the dependencies between labels. Different domains often have different label sets, which makes it difficult to directly utilize the label dependencies learned from one domain in another domain. In this paper, we introduce the dependency transfer mechanism that addresses such label-discrepancy problem. The dependency transfer mechanism learns the abstract label transition patterns from the source domains and generalizes such patterns in the target domain to benefit the prediction of a label sequence. We also develop the sequence matching network by adapting the matching network to sequence labeling case. Moreover, we propose a CRF-based few-shot sequence labeling framework to integrate both the dependency transfer mechanism and the sequence matching network. Experiments on slot tagging (ST) and named entity recognition (NER) datasets show that our model significantly outperforms the strongest few-shot learning baseline by 7.96 and 11.70 F1 scores respectively in the 1-shot setting.

Pre-Training with Whole Word Masking for Chinese BERT

Bidirectional Encoder Representations from Transformers (BERT) has shown marvelous improvements across various NLP tasks. Recently, an upgraded version of BERT has been released with Whole Word Masking (WWM), which mitigate the drawbacks of masking partial WordPiece tokens in pre-training BERT. In this technical report, we adapt whole word masking in Chinese text, that masking the whole word instead of masking Chinese characters, which could bring another challenge in Masked Language Model (MLM) pre-training task. The model was trained on the latest Chinese Wikipedia dump. We aim to provide easy extensibility and better performance for Chinese BERT without changing any neural architecture or even hyper-parameters. The model is verified on various NLP tasks, across sentence-level to document-level, including sentiment classification (ChnSentiCorp, Sina Weibo), named entity recognition (People Daily, MSRA-NER), natural language inference (XNLI), sentence pair matching (LCQMC, BQ Corpus), and machine reading comprehension (CMRC 2018, DRCD, CAIL RC). Experimental results on these datasets show that the whole word masking could bring another significant gain. Moreover, we also examine the effectiveness of Chinese pre-trained models: BERT, ERNIE, BERT-wwm. We release the pre-trained model (both TensorFlow and PyTorch) on GitHub: https://github.com/ymcui/Chinese-BERT-wwm

Learning to Ask Unanswerable Questions for Machine Reading Comprehension

Machine reading comprehension with unanswerable questions is a challenging task. In this work, we propose a data augmentation technique by automatically generating relevant unanswerable questions according to an answerable question paired with its corresponding paragraph that contains the answer. We introduce a pair-to-sequence model for unanswerable question generation, which effectively captures the interactions between the question and the paragraph. We also present a way to construct training data for our question generation models by leveraging the existing reading comprehension dataset. Experimental results show that the pair-to-sequence model performs consistently better compared with the sequence-to-sequence baseline. We further use the automatically generated unanswerable questions as a means of data augmentation on the SQuAD 2.0 dataset, yielding 1.9 absolute F1 improvement with BERT-base model and 1.7 absolute F1 improvement with BERT-large model.