LY Corporation R&D

Our papers have been accepted at The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS 2026)(external link)

Nearly-Optimal Algorithm for Adversarial Kernelized Bandits Shogo Iwazaki

Where and When Identity Forms: Identity-Vital Attention Redistribution for Training-Free Subject-Driven Generation Luan Thanh Trinh, Kenji Doi, Atsuki Osanai

How Neural Reward Models Learn Features for Policy Optimization: A Single-Index Analysis Rei Higuchi (University of Tokyo/RIKEN AIP), Ryotaro Kawata (University of Tokyo/RIKEN AIP), Akifumi Wachi, Shokichi Takakura, Kohei Miyaguchi, Taiji Suzuki (University of Tokyo/RIKEN AIP)

Reward-Estimated Hypergradient for Bilevel Reinforcement Learning with Black-Box Follower Shigeki Kusaka (University of Tsukuba), Mikoto Kudo (University of Tsukuba/RIKEN AIP), Takumi Tanabe, Akifumi Wachi, Youhei Akimoto (University of Tsukuba/RIKEN AIP)

JaSpec-LID: Local Intrinsic Dimension Estimation via Jacobian Spectra of ODE-based Generative Models Hiro Ishii (Institute of Science Tokyo), Genki Osada