no code implementations • 31 May 2023 • Yang Yang, Jian Wu, Xiangman Song, Derun Wu, Lijie Su, Lixin Tang
However, optimizing hit rate is a non-convex and challenging problem.
no code implementations • 8 Aug 2022 • Wei Luo, Tongzhi Niu, Lixin Tang, Wenyong Yu, Bin Li
At first, we propose a novel clear memory-augmented module (CMAM), which combines the encoding and memoryencoding in a way of forgetting and inputting, thereby repairing abnormal foregrounds and preserving clear backgrounds.
no code implementations • 30 Sep 2019 • Ivan Y. Tyukin, Alexander N. Gorban, Alistair A. McEwan, Sepehr Meshkinfamfard, Lixin Tang
Another feature of this approach is that, in the supervised setting, the computational complexity of training is linear in the number of training samples.