J
Jiangchao Yao
Researcher at Alibaba Group
Publications - 64
Citations - 978
Jiangchao Yao is an academic researcher from Alibaba Group. The author has contributed to research in topics: Computer science & Recommender system. The author has an hindex of 11, co-authored 47 publications receiving 663 citations. Previous affiliations of Jiangchao Yao include Shanghai Jiao Tong University.
Papers
More filters
Posted Content
How does Disagreement Help Generalization against Label Corruption
TL;DR: Co-teaching+ as discussed by the authors proposes a robust learning paradigm, which bridges the "Update by Disagreement" strategy with the original Co-Teaching, where two networks feed forward and predict all data, but keep prediction disagreement data only.
Proceedings Article
Masking: A New Perspective of Noisy Supervision
TL;DR: In this paper, a human-assisted approach called ''Masking'' is proposed that conveys human cognition of invalid class transitions and naturally speculates the structure of the noise transition matrix, which can improve the robustness of classifiers significantly.
Proceedings Article
How does disagreement help generalization against label corruption
TL;DR: Co-teaching+ as mentioned in this paper proposes a robust learning paradigm, which bridges the "Update by Disagreement" strategy with the original Co-Teaching, where two networks feed forward and predict all data, but keep prediction disagreement data only.
Journal ArticleDOI
Deep Learning From Noisy Image Labels With Quality Embedding
TL;DR: A probabilistic model is proposed, which explicitly introduces an extra variable to represent the trustworthiness of noisy labels, termed as the quality variable, which effectively minimizes the influence of label noise and outperforms the state-of-the-art deep learning approaches.
Posted Content
Masking: A New Perspective of Noisy Supervision
TL;DR: A human-assisted approach called Masking is proposed that conveys human cognition of invalid class transitions and naturally speculates the structure of the noise transition matrix and can improve the robustness of classifiers significantly.