HAR-DoReMi: Optimizing data mixture for self-supervised human activity recognition across heterogeneous IMU datasets.
Saved in:
| Title: | HAR-DoReMi: Optimizing data mixture for self-supervised human activity recognition across heterogeneous IMU datasets. |
|---|---|
| Authors: | Ban, Lulu1, llban@stu.usc.edu.cn, Zhu, Tao1, tzhu@usc.edu.cn, Lu, Xiangqing1, xqlu@stu.usc.edu.cn, Qiu, Qi1, qiuqi@stu.usc.edu.cn, Han, Wenyong1, wyhan@stu.usc.edu.cn, Li, Shuangjian2, shuangjianli@mail.dlut.edu.cn, Chen, Liming2, limingchen0922@dlut.edu.cn, Wang, Kevin I-Kai3, kevin.wang@auckland.ac.nz, Nie, Mingxing1, niemx@usc.edu.cn, Wan, Yaping1, ypwan@aliyun.com |
| Source: | Neurocomputing; Sep2026, Vol. 694, pN.PAG-N.PAG, 1p |
| Database: | Applied Science & Technology Source |
| ISSN: | 09252312 |
|---|---|
| DOI: | 10.1016/j.neucom.2026.133963 |
