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

Add to Chrome

Add to Firefox

Picture for Kenji Kawaguchi

Kenji Kawaguchi

Robustness Implies Generalization via Data-Dependent Generalization Bounds

Jun 27, 2022
Kenji Kawaguchi, Zhun Deng, Kyle Luh, Jiaoyang Huang

Figure 1 for Robustness Implies Generalization via Data-Dependent Generalization Bounds

Figure 2 for Robustness Implies Generalization via Data-Dependent Generalization Bounds

Figure 3 for Robustness Implies Generalization via Data-Dependent Generalization Bounds

Figure 4 for Robustness Implies Generalization via Data-Dependent Generalization Bounds

* Accepted by ICML 2022, and selected for ICML long presentation (top 2% of submissions)

Access Paper or Ask Questions

Set-based Meta-Interpolation for Few-Task Meta-Learning

May 20, 2022
Seanie Lee, Bruno Andreis, Kenji Kawaguchi, Juho Lee, Sung Ju Hwang

Figure 1 for Set-based Meta-Interpolation for Few-Task Meta-Learning

Figure 2 for Set-based Meta-Interpolation for Few-Task Meta-Learning

Figure 3 for Set-based Meta-Interpolation for Few-Task Meta-Learning

Figure 4 for Set-based Meta-Interpolation for Few-Task Meta-Learning

* First two authors contributed equally. Name order decided by a coin toss

Access Paper or Ask Questions

Simplicial Embeddings in Self-Supervised Learning and Downstream Classification

Apr 01, 2022
Samuel Lavoie, Christos Tsirigotis, Max Schwarzer, Kenji Kawaguchi, Ankit Vani, Aaron Courville

Figure 1 for Simplicial Embeddings in Self-Supervised Learning and Downstream Classification

Figure 2 for Simplicial Embeddings in Self-Supervised Learning and Downstream Classification

Figure 3 for Simplicial Embeddings in Self-Supervised Learning and Downstream Classification

Figure 4 for Simplicial Embeddings in Self-Supervised Learning and Downstream Classification

* 22 pages, 5 figures, 5 tables, Preprint

Access Paper or Ask Questions

EIGNN: Efficient Infinite-Depth Graph Neural Networks

Feb 22, 2022
Juncheng Liu, Kenji Kawaguchi, Bryan Hooi, Yiwei Wang, Xiaokui Xiao

Figure 1 for EIGNN: Efficient Infinite-Depth Graph Neural Networks

Figure 2 for EIGNN: Efficient Infinite-Depth Graph Neural Networks

Figure 3 for EIGNN: Efficient Infinite-Depth Graph Neural Networks

Figure 4 for EIGNN: Efficient Infinite-Depth Graph Neural Networks

* NeurIPS 2021 Camera Ready

Access Paper or Ask Questions

Adaptive Discrete Communication Bottlenecks with Dynamic Vector Quantization

Feb 02, 2022
Dianbo Liu, Alex Lamb, Xu Ji, Pascal Notsawo, Mike Mozer, Yoshua Bengio, Kenji Kawaguchi

Access Paper or Ask Questions

ExpertNet: A Symbiosis of Classification and Clustering

Jan 17, 2022
Shivin Srivastava, Kenji Kawaguchi, Vaibhav Rajan

Figure 1 for ExpertNet: A Symbiosis of Classification and Clustering

Figure 2 for ExpertNet: A Symbiosis of Classification and Clustering

Figure 3 for ExpertNet: A Symbiosis of Classification and Clustering

Figure 4 for ExpertNet: A Symbiosis of Classification and Clustering

* 16 pages, 3 figures

Access Paper or Ask Questions

Training Free Graph Neural Networks for Graph Matching

Jan 14, 2022
Zhiyuan Liu, Yixin Cao, Fuli Feng, Xiang Wang, Xindi Shang, Jie Tang, Kenji Kawaguchi, Tat-Seng Chua

Figure 1 for Training Free Graph Neural Networks for Graph Matching

Figure 2 for Training Free Graph Neural Networks for Graph Matching

Figure 3 for Training Free Graph Neural Networks for Graph Matching

Figure 4 for Training Free Graph Neural Networks for Graph Matching

* Main paper 11 papes

Access Paper or Ask Questions

Noether Networks: Meta-Learning Useful Conserved Quantities

Dec 06, 2021
Ferran Alet, Dylan Doblar, Allan Zhou, Joshua Tenenbaum, Kenji Kawaguchi, Chelsea Finn

Figure 1 for Noether Networks: Meta-Learning Useful Conserved Quantities

Figure 2 for Noether Networks: Meta-Learning Useful Conserved Quantities

Figure 3 for Noether Networks: Meta-Learning Useful Conserved Quantities

Figure 4 for Noether Networks: Meta-Learning Useful Conserved Quantities

* Accepted to NeurIPS '21. The first two authors contributed equally

Access Paper or Ask Questions