no code implementations • ICCV 2023 • Chang Zou, Zeqi Chen, Zhichao Cui, Yuehu Liu, Chi Zhang
To completely and accurately represent the information contained in a cluster and learn discriminative features, we propose to maintain discrepant cluster proxies and multi-instance proxies for a cluster.
no code implementations • 7 Oct 2022 • Zeqi Chen, Zhichao Cui, Chi Zhang, Jiahuan Zhou, Yuehu Liu
However, training two networks with a set of noisy pseudo labels reduces the complementarity of the two networks and results in label noise accumulation.