| Presentation | 2023-09-21 Analysis of subtasks for improving the detection accuracy of offensive tweets in multitask learning Ryoichi Sawada, Yu Suzuki, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | There are studies on detecting offensive tweets, but there is a need to further improve the accuracy.One method to improve accuracy is multitask learning.Multitask learning is a method to improve generalization performance by solving multiple related tasks with a model.There are some studies on combining multitask learning with the offensive tweets detection task.However, it is not obvious which subtasks are effective in improving the accuracy of offensive tweets detection.Therefore, in this study, we proposed a task that improves the accuracy of the offensive tweet detection task.We also conducted experiments to verify whether the proposed task is effective in improving the accuracy of the offensive tweet detection task.Experimental results showed that solving the anger emotion detection task simultaneously with the offensive tweet detection task using multitask learning was more effective than solving the random task simultaneously.In addition, the weighted loss function was found to be effective in improving the F value of the simultaneously solved detection task only when the number of data in the simultaneously solved detection task was unbalanced and the F value was very low. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | offensive tweet / weighted loss / multitask learning |
| Paper # | DE2023-14 |
| Date of Issue | 2023-09-14 (DE) |
| Conference Information | |
| Committee | DE / IPSJ-DBS / IPSJ-IFAT |
|---|---|
| Conference Date | 2023/9/21(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Kitakyushu International Conference Center |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Bigdata management, information retrieval, knowledge discovery, etc. |
| Chair | Masashi Toyoda(Univ. of Tokyo) |
| Vice Chair | Kosuke Takano(Kanagawa Inst. of Tech.) / Chiemi Watanabe(Tsukuba Univ. of Technology) |
| Secretary | Kosuke Takano(Univ. of Tsukuba) / Chiemi Watanabe(Komazawa Univ.) |
| Assistant | Takahiro Komamizu(Nagoya Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Data Engineering / Special Interest Group on Database System / Special Interest Group on Information Fundamentals and Access Technologies |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Analysis of subtasks for improving the detection accuracy of offensive tweets in multitask learning |
| Sub Title (in English) | |
| Keyword(1) | offensive tweet |
| Keyword(2) | weighted loss |
| Keyword(3) | multitask learning |
| 1st Author's Name | Ryoichi Sawada |
| 1st Author's Affiliation | Gifu University(Gifu) |
| 2nd Author's Name | Yu Suzuki |
| 2nd Author's Affiliation | Gifu University(Gifu) |
| Date | 2023-09-21 |
| Paper # | DE2023-14 |
| Volume (vol) | vol.123 |
| Number (no) | DE-192 |
| Page | pp.pp.19-24(DE), |
| #Pages | 6 |
| Date of Issue | 2023-09-14 (DE) |