Y
Yue He
Researcher at Tsinghua University
Publications - 15
Citations - 371
Yue He is an academic researcher from Tsinghua University. The author has contributed to research in topics: Generalization & Computer science. The author has an hindex of 6, co-authored 11 publications receiving 63 citations. Previous affiliations of Yue He include Beihang University.
Papers
More filters
Proceedings ArticleDOI
Deep Stable Learning for Out-Of-Distribution Generalization
TL;DR: In this paper, the authors propose to remove the dependencies between features via learning weights for training samples, which helps deep models get rid of spurious correlations and, in turn, concentrate more on the true connection between discriminative features and labels.
Posted Content
Towards Non-I.I.D. Image Classification: A Dataset and Baselines.
Yue He,Zheyan Shen,Peng Cui +2 more
TL;DR: The experimental results demonstrate that NICO can well support the training of ConvNet model from scratch, and a batch balancing module can help ConvNets to perform better in Non-I.I.D.D., situations with sufficient flexibility.
Journal ArticleDOI
Towards Non-I.I.D. image classification: A dataset and baselines
Yue He,Zheyan Shen,Peng Cui +2 more
TL;DR: In this paper, a Non-I.I.D. image dataset called NICO 4, which uses contexts to create non-IIDness consciously, was constructed and released.
Proceedings Article
Counterfactual Prediction for Bundle Treatment
TL;DR: This work proposes a novel variational sample re-weighting (VSR) method to eliminate confounding bias by decorrelating the treatments and confounders and conducts extensive experiments to demonstrate that the predictive model trained on this re-weightsed dataset can achieve more accurate counterfactual outcome prediction.
Proceedings ArticleDOI
StylizedNeRF: Consistent 3D Scene Stylization as Stylized NeRF via 2D-3D Mutual Learning
TL;DR: A novel mutual learning framework for 3D scene stylization that combines a 2D image stylization network and NeRF to fuse the stylization ability of 2D stylized network with the 3D consistency of NeRF is proposed.