| Presentation | 2023-03-14 Temporal relation identification toward generating temporal logic formulas Maiko Onishi, Shinpei Ogata, Kozo Okano, Daisuke Bekki, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | There is room to utilize temporal relations in relation extraction that is incorporated in the analysis of requirements specifications. Several studies have employed rule-based relation extraction. For example, it is often incorporated into methods as a component of automatic generation of temporal logic formulas and state transition models. Since rules are created based on concrete examples, the scope of application tends to be limited. Therefore, it is necessary to be able to flexibly change the composition and performance of rule-based methods in response to increases or decreases in the scale of examples. However, it is not easy to increase the scalability and generality of methods in the current situation. On the other hand, in the field of natural language processing, deep learning-based relation identification has been developed through TimeBank, which is a corpus annotated with temporal relations. But data in requirements engineering is not sufficiently large enough to learn relational identification. In addition, few studies have utilized deep learning-based relation extraction for challenging tasks in requirements engineering. This study confirms the effectiveness of deep learning-based temporal relation identification when applied to tasks, where rule-based methods have been employed in the past. Using a news domain corpus annotated with temporal relations commonly employed in the field of natural language processing, we train a model to learn temporal relations between events in a sentence. The trained model is then used for fine tuning. Property specification patterns are a basic pattern matching method employed in the synthesis of temporal logic formulas. We applied this method to test data containing patterns that abstract requirement, and demonstrated its effectiveness. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Software Engineering / Natural Language Processing / Temporal Relation Identification / Requirements Specification |
| Paper # | SS2022-49 |
| Date of Issue | 2023-03-07 (SS) |
| Conference Information | |
| Committee | SS |
|---|---|
| Conference Date | 2023/3/14(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Kozo Okano(Shinshu Univ.) |
| Vice Chair | Yoshiki Higo(Osaka Univ.) |
| Secretary | Yoshiki Higo(Shinshu Univ.) |
| Assistant | Shinsuke Matsumoto(Osaka Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Software Science |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Temporal relation identification toward generating temporal logic formulas |
| Sub Title (in English) | |
| Keyword(1) | Software Engineering |
| Keyword(2) | Natural Language Processing |
| Keyword(3) | Temporal Relation Identification |
| Keyword(4) | Requirements Specification |
| 1st Author's Name | Maiko Onishi |
| 1st Author's Affiliation | Ochanomizu University(Ochanomizu Univ.) |
| 2nd Author's Name | Shinpei Ogata |
| 2nd Author's Affiliation | Shinshu University(Shinshu Univ.) |
| 3rd Author's Name | Kozo Okano |
| 3rd Author's Affiliation | Shinshu University(Shinshu Univ.) |
| 4th Author's Name | Daisuke Bekki |
| 4th Author's Affiliation | Ochanomizu University(Ochanomizu Univ.) |
| Date | 2023-03-14 |
| Paper # | SS2022-49 |
| Volume (vol) | vol.122 |
| Number (no) | SS-432 |
| Page | pp.pp.13-18(SS), |
| #Pages | 6 |
| Date of Issue | 2023-03-07 (SS) |