Semantic Shift Stability: Efficient Way to Detect Performance Degradation of Word Embeddings and Pre-trained Language Models

Semantic Shift Stability: Efficient Way to Detect Performance Degradation of Word Embeddings and Pre-trained Language Models Shotaro Ishihara author Hiromu Takahashi author Hono Shirai author 2022-11 text Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) Yulan He editor Heng Ji editor Sujian Li editor Yang Liu editor Chua-Hui Chang editor Association for Computational Linguistics Online only conference publication ishihara-etal-2022-semantic 10.18653/v1/2022.aacl-main.17 https://aclanthology.org/2022.aacl-main.17/ 2022-11 205 216