Futoshi Iwama

Semantic analysis technique of logics retrieval for software testing from specification documents

Software testing often targets natural language specification documents. Creating test cases depends on engineer skills, then automation of creating test cases from natural language specification is important. Logics retrieval is a required technique to automate creating test cases, because once logics are retrieved we can transform them into decision tables and also create test cases from the decision tables. Furthermore, Japanese language structure is different from English. If we target Japanese natural language, a new technique is also required. We propose a Semantic Analysis Technique of Logics Retrieval for Software Testing from Japanese Public Sectors Specification Documents. This technique is a new logics retrieval from harmonization between natural language processing technique and software testing. Applying the analysis technique to total 25 files, 1,218 pages and a million double bytes characters, the precision reached 0.93 to 0.97 and recall reached 0.65 to 0.79.

Constructing parser for industrial software specifications containing formal and natural language description

This paper describes a novel framework for creating a parser to process and analyze texts written in a “partially structured” natural language. In many projects, the contents of document artifacts tend to be described as a mixture of formal parts (i.e. the text constructs follow specific conventions) and parts written in arbitrary free text. Formal parsers, typically defined and used to process a description with rigidly defined syntax such as program source code are very precise and efficient in processing the formal part, while parsers developed for natural language processing (NLP) are good at robustly interpreting the free-text part. Therefore, combining these parsers with different characteristics can allow for more flexible and practical processing of various project documents. Unfortunately, conventional approaches to constructing a parser from multiple small parsers were studied extensively only for formal language parsers and are not directly applicable to NLP parsers due to the differences in the way the input text is extracted and evaluated. We propose a method to configure and generate a combined parser by extending an approach based on parser combinator, the operators for composing multiple formal parsers, to support both NLP and formal parsers. The resulting text parser is based on Parsing Expression Grammars, and it benefits from the strength of both parser types. We demonstrate an application of such combined parser in practical situations and show that the proposed approach can efficiently construct a parser for analyzing project-specific industrial specification documents.

Hierarchical context supplementation for consecutive question answering

Question-answering (QA) systems have recently shown impressive results in terms of accurately answering user questions in such situations as domain specific user questions. However, we have identified many real situations where QA systems must cope with not a single question-answering situation but rather a sequence of consecutive questions. In such cases, users often ask questions on the basis of the previous answer they have received, so the context of the questions changes on a certain level. The commonly used method to handle this problem when using a QA system is to append the current question to the previous question (append method). However, the append method is not designed to detect such context changes. To deal with such context changes, we have designed a hierarchical context supplementation QA System (HCSQ). The HCSQ handles consecutive questions by matching the current question with the hierarchical domain knowledge database structure of the previous answer and then supplements the context of the current question with the required keywords. We also show that our method can be further applied to the initial question to supplement omitted context. Experimental results show that our method substantially outperforms the state-of-the-art methods.