Despite advancements in information access and natural language processing technologies, research on information retrieval for non-textual information in real-world documents remains limited. Tables, in particular, serve as crucial sources of various kinds of information, which makes structuring tabular data an important issue. In this work, we focus on understanding the structures of tables in Japanese annual securities reports, a specific type of real-world document, and undertake a cell-type classification task. We constructed a new dataset by manually annotating over 111,000 cells from more than 4,000 tables. In addition to implementing a baseline program, we organized a shared task using this new dataset. The results revealed that the best-performing system completed only 75% of the tables, thus indicating the ongoing challenge of understanding table structures. Our dataset and baseline program will be made available on GitHub for researchers in this field.

A dialog system that can monitor the health status of seniors has a huge potential for solving the labor force shortage in the caregiving industry in aging societies. As a part of efforts to create such a system, we are developing two modules that are aimed to correctly interpret user utterances: (i) a yes/no response classifier, which categorizes responses to health-related yes/no questions that the system asks; and (ii) an entailment recognizer, which detects users’ voluntary mentions about their health status. To apply machine learning approaches to the development of the modules, we created large annotated datasets of 280,467 question-response pairs and 38,868 voluntary utterances. For question-response pairs, we asked annotators to avoid direct “yes” or “no” answers, so that our data could cover a wide range of possible natural language responses. The two modules were implemented by fine-tuning a BERT model, which is a recent successful neural network model. For the yes/no response classifier, the macro-average of the average precisions (APs) over all of our four categories (Yes/No/Unknown/Other) was 82.6% (96.3% for “yes” responses and 91.8% for “no” responses), while for the entailment recognizer it was 89.9%.

In this paper, we propose a method for whyquestion answering (why-QA) that uses an adversarial learning framework.Existing why-QA methods retrieve answer passages that usually consist of several sentences.These multi-sentence passages contain not only the reason sought by a why-question and its connection to the why-question, but also redundant and/or unrelated parts.We use our proposed Adversarial networks for Generating compact-answer Representation (AGR) to generate from a passage a vector representation of the non-redundant reason sought by a why-question and exploit the representation for judging whether the passage actually answers the why-question.Through a series of experiments using Japanese why-QA datasets, we show that these representations improve the performance of our why-QA neural model as well as that of a BERT-based why-QA model.We show that they also improve a state-of-the-art distantly supervised open-domain QA (DS-QA) method on publicly available English datasets, even though the target task is not a why-QA.

Kazuma Kadowaki, Ryu Iida, Kentaro Torisawa, Jong-Hoon Oh, Julien Kloetzer. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

In this paper, we proposed a user management scheme for Peer-to-Peer (P2P) contents sharing systems. The proposed scheme assumes the existence of peer groups which provide sensitive, confidential contents to the group members, and controls the membership through the collaboration among group members. As the key technique, we use the delegation of the task of qualification to few selected group members, and a voting mechanism based on the (n, t)-threshold cryptosystem. We implemented a prototype of the proposed scheme using JXTA framework, and evaluated the performance experimentally. The experimental result indicates that although the time required for the admission control is proportional to the group size and the threshold used in the threshold cryptosystem, we can significantly reduce it by executing it in parallel. We also demonstrate the availability of the scheme by applying it to the P2P live video streaming system.

In this paper, we propose a user management system for networked consumer electronics (nCE). The proposed system extends Tokunaga's hybrid P2P system in such a way that the inquiries concerned with client peers submitted by service-provider peers are handled by several authentication proxies instead of a centralized management peer, although client peers still have to register themselves to the centralized server. A prototype of the proposed system is implemented using JXTA framework, and is evaluated via experiments. The result of experiments indicates that the proposed system certainly reduces both the load of the centralized server and the time required for the authentication process.