A machine learner. NEC Corp. (2017-), RIKEN AIP(2018-2022), also an indep. researcher (2023-). Elementary particle physics (-2017), Machine learning (2017-).
Our novel DNN-based solver for functional differential equations (FDEs) reduces computational cost from ~days to ~hours and enhances expressivity from polynomials of ~10 degrees to 1000 degrees.
A
我々の論文が #ICLR2021 にSpotlight (top 6% of 3000 submitted papers!) で採択されました!
"Sequential Density Ratio Estimation for Simultaneous Optimization of Speed and Accuracy"
[海老原章記
Just released super-cool "paper-explained" video for our paper accepted at #NeurIPS2024 !!
CHECK IT OUT NOW
The media could not be played.
Quote
Taiki Miyagawa @ 論文千本ノックの人
@kanaheinousagi
·
Accepted at NeurIPS2024!!!!!#NeurIPS24#NeurIPS24#NeurIPS2024@NeurIPSConf
Our novel DNN-based solver for functional differential equations (FDEs) reduces computational cost from ~days to ~hours and enhances expressivity from polynomials of ~10 degrees to 1000 degrees.
A
Another paper has been accepted at #WACV2025 !!
“Federated source-free domain adaptation for classification: Weighted cluster aggregation for unlabeled data”
Junki Mori, Kosuke Kihara, Taiki Miyagawa, Akinori F. Ebihara, Isamu Teranishi, and Hisashi Kashima