Daichi Yashima

Robot velocity fields, not just sparse trajectories, unlock embodiment-agnostic learning for language-conditioned manipulation.

Rigel achieves over 10-point improvements in image and video captioning evaluation by aligning automatic metrics with human judgment without relying on large vocabulary sets.

Achieving 33× faster robot flow generation while improving success rates in manipulation tasks could redefine efficiency benchmarks in robotic motion planning.

ELSA outperforms existing metrics by aligning audio evaluations with human judgments at the level of individual acoustic events, transforming TTA assessment.

Turns out, your MLLM judge is probably biased: MLLMs tend to prefer their own generations and those from similar architectures, potentially skewing benchmark results.

Achieve 3x faster video captioning without sacrificing accuracy by swapping quadratic attention for a linear Mamba backbone and hierarchical bidirectional scanning.

A lightweight VLA with deep state space models lets robots outperform larger models at language-guided manipulation while running 3x faster.

Nail design retrieval gets a major upgrade: NaiLIA leverages dense intent descriptions and palette queries to outperform standard methods, opening the door to more nuanced and personalized image search.

Key contribution not extracted.