Research

Publications

Research Area

Year

2024

International Conferences

Representation Learning Theory

“ADOPT: Modified Adam Can Converge with Any β2 with the Optimal Rate”

Shohei Taniguchi, Keno Harada, Gouki Minegishi, Yuta Oshima, Seong Cheol Jeong, Go Nagahara, Tomoshi Iiyama, Masahiro Suzuki, Yusuke Iwasawa, Yutaka Matsuo

Advances in Neural Information Processing Systems 37 (NeurIPS 2024)
2024

Journals

Representation Learning 理論生成モデル

Multi-agent symbol emergence and latent representation learning by integration of Gaussian process latent variable models and neural networks

中村友昭, 鈴木雅大, 谷口彰, 谷口忠大

日本ロボット学会誌（レター）, (2024)
2023

Journals

computer-vision Representation Learning 理論

HAWK-Net: Hierarchical Attention Weighted Top-K Network for Megapixel Image Classification

Hitoshi Nakanishi, Masahiro Suzuki, Yutaka Matsuo

IPSJ, (2023).
2022

Journals

computer-vision Representation Learning

Fixing the train-test objective discrepancy: Iterative Image Inpainting for Unsupervised Anomaly Detection

Hitoshi Nakanishi, Masahiro Suzuki, Yutaka Matuo.

J-Stage in August Vol.30, (2022).
2021

International Conferences

Computer Vision Representation Learning Theory

Test-Time Classifier Adjustment Module for Model-Agnostic Domain Generalization

Yusuke Iwasawa, and Yutaka Matsuo.

Advances in Neural Information Processing Systems 2021 (NeurIPS2021, Spotlight). December 2021.
2021

International Conferences

Reinforcement Learning Representation Learning Theory

Co-Adaptation of Algorithmic and Implementational Innovations in Inference-based Deep Reinforcement Learning

Hiroki Furuta, Tadashi Kozuno, Tatsuya Matsushima, Yutaka Matsuo, and Shixiang Shane Gu.

Advances in Neural Information Processing Systems 2021 (NeurIPS2021). December 2021.
2021

Journals

Representation Learning 理論
Information-theoretic regularization for learning global features by sequential VAE
- https://s3.ap-northeast-1.amazonaws.com/publication.weblab/2021_Kei+Akuzawa_ECML+Journal+Track_Information-theoretic+regularization+for+learning+global+features+by+sequential+VAE.pdf
Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo

Mach Learn (2021)
2021

Journals

Representation Learning 理論
Out-of-distribution Detection Using Joint Probability between Class and Geometric Transformation
- https://s3.ap-northeast-1.amazonaws.com/publication.weblab/2021_%E5%B2%A1%E6%9C%AC%E5%BC%98%E9%87%8E_%E6%83%85%E5%A0%B1%E5%87%A6%E7%90%86%E5%AD%A6%E4%BC%9A%E8%AB%96%E6%96%87%E8%AA%8C_%E6%B7%B1%E5%B1%A4%E3%83%8B%E3%83%A5%E3%83%BC%E3%83%A9%E3%83%AB%E3%83%8D%E3%83%83%E3%83%88%E3%83%AF%E3%83%BC%E3%82%AF%E3%81%AB%E3%82%88%E3%82%8B%E3%82%AF%E3%83%A9%E3%82%B9%E3%81%A8%E5%B9%BE%E4%BD%95%E5%A4%89%E6%8F%9B%E3%81%AE%E5%90%8C%E6%99%82%E5%88%86%E9%A1%9E%E7%A2%BA%E7%8E%87%E3%82%92%E5%88%A9%E7%94%A8%E3%81%97%E3%81%9F%E5%88%86%E5%B8%83%E5%A4%96%E6%A4%9C%E7%9F%A5.pdf
岡本弘野, 鈴木雅大, 松尾豊

情報処理学会論文誌, Vol.62, No.7, pp.1382-1392 (2021)
2021

Journals

Representation Learning 理論
Semi-supervised Out-of-distribution Detection Using Output of Intermediate Layer in Deep Neural Networks
- https://s3.ap-northeast-1.amazonaws.com/publication.weblab/2020_%E5%B2%A1%E6%9C%AC%E5%BC%98%E9%87%8E_%E6%83%85%E5%A0%B1%E5%87%A6%E7%90%86%E5%AD%A6%E4%BC%9A%E8%AB%96%E6%96%87%E8%AA%8C_%E6%B7%B1%E5%B1%A4%E3%83%8B%E3%83%A5%E3%83%BC%E3%83%A9%E3%83%AB%E3%83%8D%E3%83%83%E3%83%88%E3%83%AF%E3%83%BC%E3%82%AF%E3%81%AE%E4%B8%AD%E9%96%93%E5%B1%A4%E5%87%BA%E5%8A%9B%E3%82%92%E5%88%A9%E7%94%A8%E3%81%97%E3%81%9F%E5%8D%8A%E6%95%99%E5%B8%AB%E3%81%82%E3%82%8A%E5%88%86%E5%B8%83%E5%A4%96%E6%A4%9C%E7%9F%A5.pdf
岡本弘野, 鈴木雅大, 松尾豊

情報処理学会論文誌, Vol.62, No.4, pp.1142-1151 (2021)
2021

International Conferences

Reinforcement Learning Representation Learning Theory

Policy Information Capacity: Information-Theoretic Measure for Task Complexity in Deep Reinforcement Learning

Hiroki Furuta, Tatsuya Matsushima, Tadashi Kozuno, Yutaka Matsuo, Sergey Levine, Ofir Nachum, and Shixiang Shane Gu

International Conference on Machine Learning 2021 (ICML2021).

Research

Publications

Category

Research Area

Year

“ADOPT: Modified Adam Can Converge with Any β2 with the Optimal Rate”

Multi-agent symbol emergence and latent representation learning by integration of Gaussian process latent variable models and neural networks

HAWK-Net: Hierarchical Attention Weighted Top-K Network for Megapixel Image Classification

Fixing the train-test objective discrepancy: Iterative Image Inpainting for Unsupervised Anomaly Detection

Test-Time Classifier Adjustment Module for Model-Agnostic Domain Generalization

Co-Adaptation of Algorithmic and Implementational Innovations in Inference-based Deep Reinforcement Learning

Information-theoretic regularization for learning global features by sequential VAE

Out-of-distribution Detection Using Joint Probability between Class and Geometric Transformation

Semi-supervised Out-of-distribution Detection Using Output of Intermediate Layer in Deep Neural Networks

Policy Information Capacity: Information-Theoretic Measure for Task Complexity in Deep Reinforcement Learning