Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Picture for Masashi Sugiyama

Mediated Uncoupled Learning: Learning Functions without Direct Input-output Correspondences


Jul 16, 2021
Ikko Yamane, Junya Honda, Florian Yger, Masashi Sugiyama

* ICML 2021 version with correction to Figure 1 

  Access Paper or Ask Questions

Positive-Unlabeled Classification under Class-Prior Shift: A Prior-invariant Approach Based on Density Ratio Estimation


Jul 11, 2021
Shota Nakajima, Masashi Sugiyama

* 18 pages, 4 figures 

  Access Paper or Ask Questions

Seeing Differently, Acting Similarly: Imitation Learning with Heterogeneous Observations


Jun 17, 2021
Xin-Qiang Cai, Yao-Xiang Ding, Zi-Xuan Chen, Yuan Jiang, Masashi Sugiyama, Zhi-Hua Zhou

* 17 pages, 25 figures 

  Access Paper or Ask Questions

Multi-Class Classification from Single-Class Data with Confidences


Jun 16, 2021
Yuzhou Cao, Lei Feng, Senlin Shu, Yitian Xu, Bo An, Gang Niu, Masashi Sugiyama

* 23 pages, 1 figure 

  Access Paper or Ask Questions

Probabilistic Margins for Instance Reweighting in Adversarial Training


Jun 15, 2021
Qizhou Wang, Feng Liu, Bo Han, Tongliang Liu, Chen Gong, Gang Niu, Mingyuan Zhou, Masashi Sugiyama

* 17 pages, 4 figures 

  Access Paper or Ask Questions

On the Robustness of Average Losses for Partial-Label Learning


Jun 11, 2021
Jiaqi Lv, Lei Feng, Miao Xu, Bo An, Gang Niu, Xin Geng, Masashi Sugiyama


  Access Paper or Ask Questions

Loss function based second-order Jensen inequality and its application to particle variational inference


Jun 10, 2021
Futoshi Futami, Tomoharu Iwata, Naonori Ueda, Issei Sato, Masashi Sugiyama


  Access Paper or Ask Questions

Instance Correction for Learning with Open-set Noisy Labels


Jun 01, 2021
Xiaobo Xia, Tongliang Liu, Bo Han, Mingming Gong, Jun Yu, Gang Niu, Masashi Sugiyama


  Access Paper or Ask Questions

Sample Selection with Uncertainty of Losses for Learning with Noisy Labels


Jun 01, 2021
Xiaobo Xia, Tongliang Liu, Bo Han, Mingming Gong, Jun Yu, Gang Niu, Masashi Sugiyama


  Access Paper or Ask Questions

A unified view of likelihood ratio and reparameterization gradients


May 31, 2021
Paavo Parmas, Masashi Sugiyama

* In International Conference on Artificial Intelligence and Statistics (pp. 4078-4086). PMLR (2021, March) 
* AISTATS2021; Earlier paper was split in two (arXiv:1910.06419). Refer to the current paper for the unified view, but see the earlier paper for discussion on an importance sampling technique 

  Access Paper or Ask Questions

NoiLIn: Do Noisy Labels Always Hurt Adversarial Training?


May 31, 2021
Jingfeng Zhang, Xilie Xu, Bo Han, Tongliang Liu, Gang Niu, Lizhen Cui, Masashi Sugiyama


  Access Paper or Ask Questions

Positive-Negative Momentum: Manipulating Stochastic Gradient Noise to Improve Generalization


Mar 31, 2021
Zeke Xie, Li Yuan, Zhanxing Zhu, Masashi Sugiyama

* 19 pages, 12 figures 

  Access Paper or Ask Questions

Approximating Instance-Dependent Noise via Instance-Confidence Embedding


Mar 25, 2021
Yivan Zhang, Masashi Sugiyama


  Access Paper or Ask Questions

Discovering Diverse Solutions in Deep Reinforcement Learning


Mar 12, 2021
Takayuki Osa, Voot Tangkaratt, Masashi Sugiyama

* 18 pages 

  Access Paper or Ask Questions

Lower-bounded proper losses for weakly supervised classification


Mar 04, 2021
Shuhei M. Yoshida, Takashi Takenouchi, Masashi Sugiyama


  Access Paper or Ask Questions

LocalDrop: A Hybrid Regularization for Deep Neural Networks


Mar 01, 2021
Ziqing Lu, Chang Xu, Bo Du, Takashi Ishida, Lefei Zhang, Masashi Sugiyama


  Access Paper or Ask Questions

Incorporating Causal Graphical Prior Knowledge into Predictive Modeling via Simple Data Augmentation


Feb 27, 2021
Takeshi Teshima, Masashi Sugiyama

* 24 pages, 5 figures, 2 tables 

  Access Paper or Ask Questions

Guided Interpolation for Adversarial Training


Feb 15, 2021
Chen Chen, Jingfeng Zhang, Xilie Xu, Tianlei Hu, Gang Niu, Gang Chen, Masashi Sugiyama


  Access Paper or Ask Questions

Learning from Similarity-Confidence Data


Feb 13, 2021
Yuzhou Cao, Lei Feng, Yitian Xu, Bo An, Gang Niu, Masashi Sugiyama

* 33 pages, 5 figures 

  Access Paper or Ask Questions

CIFS: Improving Adversarial Robustness of CNNs via Channel-wise Importance-based Feature Selection


Feb 10, 2021
Hanshu Yan, Jingfeng Zhang, Gang Niu, Jiashi Feng, Vincent Y. F. Tan, Masashi Sugiyama


  Access Paper or Ask Questions

Understanding the Interaction of Adversarial Training with Noisy Labels


Feb 09, 2021
Jianing Zhu, Jingfeng Zhang, Bo Han, Tongliang Liu, Gang Niu, Hongxia Yang, Mohan Kankanhalli, Masashi Sugiyama


  Access Paper or Ask Questions

Learning Diverse-Structured Networks for Adversarial Robustness


Feb 08, 2021
Xuefeng Du, Jingfeng Zhang, Bo Han, Tongliang Liu, Yu Rong, Gang Niu, Junzhou Huang, Masashi Sugiyama

* 26 pages, 8 figures 

  Access Paper or Ask Questions

Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization


Feb 04, 2021
Yivan Zhang, Gang Niu, Masashi Sugiyama


  Access Paper or Ask Questions

Provably End-to-end Label-Noise Learning without Anchor Points


Feb 04, 2021
Xuefeng Li, Tongliang Liu, Bo Han, Gang Niu, Masashi Sugiyama


  Access Paper or Ask Questions

Binary Classification from Multiple Unlabeled Datasets via Surrogate Set Classification


Feb 01, 2021
Shida Lei, Nan Lu, Gang Niu, Issei Sato, Masashi Sugiyama


  Access Paper or Ask Questions

Source-free Domain Adaptation via Distributional Alignment by Matching Batch Normalization Statistics


Jan 19, 2021
Masato Ishii, Masashi Sugiyama


  Access Paper or Ask Questions

A Symmetric Loss Perspective of Reliable Machine Learning


Jan 05, 2021
Nontawat Charoenphakdee, Jongyeong Lee, Masashi Sugiyama

* Preprint of an Invited Review Article 

  Access Paper or Ask Questions

Combinatorial Pure Exploration with Full-bandit Feedback and Beyond: Solving Combinatorial Optimization under Uncertainty with Limited Observation


Dec 31, 2020
Yuko Kuroki, Junya Honda, Masashi Sugiyama

* Preprint of an Invited Review Article 

  Access Paper or Ask Questions