The idea of hierarchical index is applied to the legal domain to provide the readers a general understanding of legal concepts via their super/sub-ordinate relations. This work serves as effort in automatic legal ontology learning in which super/sub-ordinate relations are considered. Indices are extracted from legal documents as keywords and their relationships are discovered by language processing method. We propose an approach to extract the super/sub- ordinate relation between each pair of concepts individually based on directional similarity. The relations among a set of legal indices are represented in a directed graph and the hierarchical structure of indices is simply exported from this graph. We adopt this proposal to the Japanese National Pension Act document. The resulted hierarchical structure is compared to an annotated legal ontology on the number of correct relations. The proposed method achieves 40.6% for precision, 46.9% for recall and 43.5% for F-measure as the performance.

Reference resolution is an important task which supports understanding natural language texts, especially in the legal domain, where legal articles are usually long and complicated. This paper focuses on the task of reference resolution in the legal domain, in which we extract references and resolve them to the referenced texts. We propose a four-step framework to deal with the task: mention detection, contextual information extraction, antecedent candidate extraction, and antecedent determination. We also show how machine learning methods can be exploited in each step. The final system achieves 80.06% in the F1 score for detecting references, 85.61% accuracy for resolving them, and 67.02% in the F1 score on the end-to-end setting task on the Japanese National Pension Law corpus. Our work provides promising results for further studies on this interesting task.

This study proposes a novel unsupervised approach for extracting keywords from Japanese legal documents by applying knowledge of Japanese syntax. Japanese keywords usually occur in chunks; the task of extracting Japanese keywords is treated as a matter of finding chunks that yield documents' important content. To find these chunks, all chunks in a given document are assigned weights to indicate their importance. Highly weighted chunks are recognized as candidate keywords, which are post-processed to obtain keywords. Although the proposed method employs simple techniques, the experimental results on Japanese legal documents show that the proposed chunk-based approach achieves better performance (10.5% higher on F1-score) than the graph-based ranking approach, the most popular unsupervised method.

Part-of-speech (POS) tagging is a fundamental task in natural language processing (NLP). It provides useful information for many other NLP tasks, including word sense disambiguation, text chunking, named entity recognition, syntactic parsing, semantic role labeling, and semantic parsing. In this paper, we present a new method for Vietnamese POS tagging using dual decomposition. We show how dual decomposition can be used to integrate a word-based model and a syllable-based model to yield a more powerful model for tagging Vietnamese sentences. We also describe experiments on the Viet Treebank corpus, a large annotated corpus for Vietnamese POS tagging. Experimental results show that our model using dual decomposition outperforms both word-based and syllable-based models.

Analyzing logical structures of texts is important to understanding natural language, especially in the legal domain, where legal texts have their own specific characteristics. Recognizing logical structures in legal texts does not only help people in understanding legal documents, but also in supporting other tasks in legal text processing. In this article, we present a new task, learning logical structures of paragraphs in legal articles, which is studied in research on Legal Engineering. The goals of this task are recognizing logical parts of law sentences in a paragraph, and then grouping related logical parts into some logical structures of formulas, which describe logical relations between logical parts. We present a two-phase framework to learn logical structures of paragraphs in legal articles. In the first phase, we model the problem of recognizing logical parts in law sentences as a multi-layer sequence learning problem, and present a CRF-based model to recognize them. In the second phase, we propose a graph-based method to group logical parts into logical structures. We consider the problem of finding a subset of complete subgraphs in a weighted-edge complete graph, where each node corresponds to a logical part, and a complete subgraph corresponds to a logical structure. We also present an integer linear programming formulation for this optimization problem. Our models achieve 74.37% in recognizing logical parts, 80.08% in recognizing logical structures, and 58.36% in the whole task on the Japanese National Pension Law corpus. Our work provides promising results for further research on this interesting task.

Recognizing Textual Entailment (RTE) is a fundamental task in Natural Language Understanding. The task is to decide whether the meaning of a text can be inferred from the meaning of another one. In this article, we conduct an empirical study of recognizing textual entailment in Japanese texts, in which we adopt a machine learning-based approach to the task. We quantitatively analyze the effects of various entailment features, machine learning algorithms, and the impact of RTE resources on the performance of an RTE system. This article also investigates the use of machine translation for the RTE task and determines whether machine translation can be used to improve the performance of our RTE system. Experimental results achieved on benchmark data sets show that our machine learning-based RTE system outperforms the baseline methods based on lexical matching and syntactic matching. The results also suggest that the machine translation component can be utilized to improve the performance of the RTE system.

Our research addresses the task of updating legal documents when new information emerges. In this paper, we employ a hierarchical ranking model to the task of updating legal documents. Word clustering features are incorporated to the ranking models to exploit semantic relations between words. Experimental results on legal data built from the United States Code show that the hierarchical ranking model with word clustering outperforms baseline methods using Vector Space Model, and word cluster-based features are effective features for the task.

A logical formulation system functions to verify consistency of legal documents, and to eliminate inconsistent parts from a set of articles. We are studying legal document analysis methods, and in this paper we focus on how to deal with multiple sentences which constitute a paragraph of an article in a law. They need to be processed together because they are semantically dependent on each other. We analyzed National Pension Law of Japan and found that relations between sentences and their logical structures can be classified into four main types. We implemented a logical formulation system, which showed reasonable accuracy in dealing properly with paragraphs consisting of multiple sentences.

We formulate semantic parsing as a parsing problem on a synchronous context free grammar (SCFG) which is automatically built on the corpus of natural language sentences and the representation of semantic outputs. We then present an online learning framework for estimating the synchronous SCFG grammar. In addition, our online learning methods for semantic parsing problems are also extended to deal with the case, in which the semantic representation could be represented under lambda-calculus. Experimental results in the domain of semantic parsing show advantages in comparison with previous works.

Sentence reduction is the removal of redundant words or phrases from an input sentence by creating a new sentence in which the gist of the original meaning of the sentence remains unchanged. All previous methods required a syntax parser before sentences could be reduced; hence it was difficult to apply them to a language with no reliable parser. In this article we propose two new sentence-reduction algorithms that do not use syntactic parsing for the input sentence. The first algorithm, based on the template-translation learning algorithm, one of example-based machine-translation methods, works quite well in reducing sentences, but its computational complexity can be exponential in certain cases. The second algorithm, an extension of the template--translation algorithm via innovative employment of the Hidden Markov model, which uses the set of template rules learned from examples, can overcome this computation problem. Experiments show that the proposed algorithms achieve acceptable results in comparison to sentence reduction done by humans.

We propose a model, called a deliberation structure model, that represents streams of relevant messages in email communication. The deliberation structure model treats e‐mail messages that include several topics and uses tree structure, called deliberation trees, that are based on Clark's contribution trees to represent message structure. Deliberation trees are useful to visualize the message structure. A method for constructing the deliberation structure is showed.

Japanese has many noun phrase patterns of the type A no B consisting of two nouns A and B with an adnominal particle no. As the semantic relations between the two nouns in the noun phrase are not made explicit, the interpretation of the phrases depends mainly on the semantic characteristics of the nouns. This paper describes the semantic diversity of A no B and a method of semantic analysis for such phrases based on feature unification.

This paper proposes an analysis method for Japanese modality. In this purpose, meaning of Japanese modality is classified into four semantic categories and the role of it is formalized into five modality functions. Based on these formalizations, information and constraints to be applied to the modality analysis procedure are specified. Then by combining these investigations with case analysis, the analysis method is proposed. This analysis method has been applied to Japanese analysis for machine translation.