denkiwakame

Working Experiences

  1. Senior Researcher at OMRON SINIC X,

    Mobile Robot Navigation / 3D Computer Vision / GPU-accelerated Machine Learning

  2. Research Engineer at OMRON SINIC X,

    Mobile Robot Navigation / 3D Computer Vision / GPU-accelerated Machine Learning

  3. Senior Software Engineer at Fixstars,

    Computer Vision / Autonomous Driving / SIMD / CUDA

  4. Researcher at NTT Labs,

    Computer Vision / Low-level Vision / MRF Optimization

  5. Part-time Web Application Engineer at Hatena,

    Frontend / Backend / Web Application

  6. Research Intern at IBM Tokyo Research Laboratory (TRL),

    Computer Vision / Traffic Monitoring

  7. Software Engineer Intern (BOLD) at Google,

    Native Client Project, Chrome Team

Education

  1. Ph.D. in Informatics at Kyoto University,

  2. M.S. in Informatics at Kyoto University,

  3. B.E. in Electrical and Electronic Engineering at Kyoto University,

Publications

  1. CLAP: Direct VLM-to-VLA Adaptation via Language-Action Grounding

    Yuri Ishitoya†, Jeremy Siburian†, Masashi Hamaya, Haruto Suzuki, Kuniaki Saito, Toshihiko Fukushima, Cristian C. Beltran-Hernandez, Mai Nishimura†, (†equal contribution)

    Conference on Robot Learning (CoRL) ,

    paper project code

  2. PHASE: Compliance-Enabled Tactile Phase Retrieval for Few-Shot Insertion Learning

    Jeremy Siburian, Cristian C. Beltran-Hernandez, Tatsuya Matsushima, Yusuke Iwasawa, Masashi Hamaya, Mai Nishimura

    International Conference on Intelligent Robots and Systems (IROS) ,

    project

  3. Can Compact Language Models Search Like Agents? Distillation-Guided Policy Optimization for Preserving Agentic RAG Capabilities

    Rikuto Kotoge, Mai Nishimura, Jiaxin Ma

    Association for Computational Linguistics (ACL) ,

    paper project

  4. Tactile Memory with Soft Robot: Robust Object insertion via Masked Encoding and Soft Wrist

    Tatusya Kamijo†, Mai Nishimura†, Nodoka Shibasaki, Jeremy Siburian, Cristian C. Beltran-Hernandez, Masahi Hamaya, (†equal contribution)

    IEEE Robotics and Automation Letters (RA-L) ,

    paper

  5. Fluidic robot swarm control for obstacle-unaware navigation

    Michikuni Eguchi, Mai Nishimura, Shigeo Yoshida, Takefumi Hiraki

    Advanced Robotics ,

    paper

  6. Multi-Agent Behavior Retrieval: Retrieval-Augmented Policy Training for Cooperative Manipulation by Mobile Robots

    So Kuroki, Mai Nishimura, Tadashi Kozuno

    International Conference on Intelligent Robots and Systems (IROS) ,

    paper project

  7. Robot Swarm Control Based on Smoothed Particle Hydrodynamics for Obstacle-Unaware Navigation

    Michikuni Eguchi, Mai Nishimura, Shigeo Yoshida, Takefumi Hiraki

    International Conference on Intelligent Robots and Systems (IROS) ,

    paper

  8. Swarm Body: Embodied Swarm Robots

    Sosuke Ichihashi, So Kuroki, Mai Nishimura, Kazumi Kasaura, Takefumi Hiraki, Kazutoshi Tanaka, Shigeo Yoshida

    ACM Conference on Human Factors in Computing Systems (CHI) ,

    paper arxiv

  9. When to Replan? An Adaptive Replanning Strategy for Autonomous Navigation using Deep Reinforcement Learning

    Kohei Honda, Ryo Yonetani, Mai Nishimura, Tadashi Kozuno

    International Conference on Robotics and Automation (ICRA) ,

    paper project

  10. Counterfactual Fairness Filter fo Fair-Delay Multi-Robot Navigation

    Hikaru Asano, Ryo Yonetani, Mai Nishimura, Tadashi Kozuno

    International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS) ,

    paper project

  11. View Birdification in the Crowd: Ground-Plane Localization from Perceived Movements

    Mai Nishimura, Shohei Nobuhara, Ko Nishino

    International Journal of Computer Vision (IJCV) ,

    paper arxiv

  12. Periodic Multi-Agent Path Planning

    Kazumi Kasaura, Ryo Yonetani, Mai Nishimura

    AAAI Conference on Artificial Intelligence (AAAI) ,

    paper project

  13. ViewBirdiformer: Learning to recover ground-plane crowd trajectories and ego-motion from a single ego-centric view

    Mai Nishimura, Shohei Nobuhara, Ko Nishino

    IEEE Robotics and Automation Letters (RA-L) ,

    paper

  14. Prioritized Safe Interval Path Planning for Multi-Agent Pathfinding with Continuous Time on 2D Roadmaps

    Kazumi Kasaura, Mai Nishimura, Ryo Yonetani

    IEEE Robotics and Automation Letters (RA-L) ,

    paper project

  15. CTRMs: Learning to Construct Cooperative Timed Roadmaps for Multi-agent Path Planning in Continuous Spaces

    Keisuke Okumura, Ryo Yonetani, Mai Nishimura, Asako Kanezaki

    International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS) ,

    paper project

  16. Edge Computing-Assisted DNN Image Recognition System With Progressive Image Retransmission

    Mutsuki Nakahara, Mai Nishimura, Yoshitaka Ushiku, Takayuki Nishio, Kazuki Maruta, Yu Nakayama, Daisuke Hisano

    IEEE Access ,

    paper

  17. View Birdification in the Crowd: Ground-Plane Localization from Perceived Movements

    Mai Nishimura, Shohei Nobuhara, Ko Nishino

    The British Machine Vision Conference (BMVC) ,

    paper

  18. Path Planning using Neural A* Search

    Ryo Yonetani, Tatsunori Taniai, Mohammadamin Barekatain, Mai Nishimura, Asako Kanezaki

    International Conference on Machine Learning (ICML) ,

    paper project

  19. Retransmission Edge Computing System Conducting Adaptive Image Compression Based on Image Recognition Accuracy

    Mutsuki Nakahara, Daisuke Hisano, Mai Nishimura, Yoshitaka Ushiku, Kazuki Maruta, Yu Nakayama

    VTC-Fall ,

    paper

  20. Crowd Density Forecasting by Modeling Patch-based Dynamics

    Hiroaki Minoura, Ryo Yonetani, Mai Nishimura, Yoshitaka Ushiku

    IEEE Robotics and Automation Letters (RA-L) ,

    paper

  21. L2B: Learning to Balance the Safety-Efficiency Trade-off in Interactive Crowd-aware Robot Navigation

    Mai Nishimura, Ryo Yonetani

    International Conference on Intelligent Robots and Systems (IROS) ,

    paper project

  22. Fast Algorithm for the Lasso based L1-Graph Construction

    Yasuhiro Fujiwara, Yasutoshi Ida, Juny Arai, Mai Nishimura, Sotetsu Iwamura

    Proceedings of the Very Large Data Bases (PVLDB) ,

    paper

  23. A Linear Generalized Camera Calibration from Three Intersecting Reference Planes

    Mai Nishimura, Shohei Nobuhara, Takashi Matsuyama, Shinya Shimizu, Kensaku Fujii

    International Conference on Computer Vision (ICCV) ,

    paper code

Workshop Papers and Extended Abstracts (Peer-Reviewed)

  1. Democratizing Agentic RAG: Distillation-Guided Policy Optimization for Compact Language Models

    Rikuto Kotoge, Mai Nishimura, Jiaxin Ma

    NeurIPS Workshop on LAW: Bridging Language, Agent, and World Models ,

    openreview workshop

  2. Tactile Memory with Soft Robot: Tactile Retrieval-based Contact-rich Manipulation with a Soft Wrist

    Tatsuya Kamijo, Mai Nishimura, Cristian Camilo Beltran-Hernandez, Masashi Hamaya

    CoRL Workshop on Dexterous Manipulation: Learning and Control with Diverse Modalities ,

    openreview

  3. FluidicSwarm: Embodiment of Swarm Robots Using Fluid Behavior Imitation

    Michikuni Eguchi, Shigeo Yoshida, Mai Nishimura, Takefumi Hiraki

    ACM SIGGRAPH 2025 Emerging Technologies (SIGGRAPH 2025) ,

    paper project

Preprints

  1. On the Effectiveness of Graph Reordering for Accelerating Approximate Nearest Neighbor Search on GPU

    Yutaro Oguri, Mai Nishimura, Yusuke Matsui

    arXiv ,

    paper

  2. InCrowdFormer: On-Ground Pedestrian World Model From Egocentric Views

    Mai Nishimura, Shohei Nobuhara, Ko Nishino

    arXiv ,

    paper

  3. TransPoser: Transformer as an Optimizer for Joint Object Shape and Pose Estimation

    Yuta Yoshitake, Mai Nishimura, Shohei Nobuhara, Ko Nishino

    arXiv ,

    paper

Pet Projects

arxiv2notion

Chrome Extension • Notion API

academic-project-template

React • UIKit

nanokit

Pixi • zsh • sheldon • starship • dotter • neovim

numpy-blas-bench

Numpy • MKL • OpenBLAS • AOCL • conda • pypi

pixi-tex

TeX • Pixi • Tectonic • XeTeX

py-tiny-pkg

Python Packaging • PEP links for my personal use

Posts

Pixi Wiki

Pixi • CUDA • docker • pypi • conda • uv

How to Report Progress Effectively

Research

Essential Pointers for Smooth NVIDIA Driver, CUDA, and Docker Integration

CUDA • docker • driver

Setting up a faiss build environment with Pixi

CUDA • Pixi • faiss • conda • PyPI

How to switch BLAS backends in numpy

Pixi • faiss • numpy • conda-forge • OpenBLAS • MKL

Manageng Dotfiles by Rust Tools

Rust • Pixi • dotfiles • nvim

Research Grants

  1. Japan Science and Technology Agency (JST), PREST :: PI

    2025- "経験の蓄積と想起を司る自律型記憶基盤"

  1. Japan Science and Technology Agency (JST), AIP Acceleration Research :: Co-PI

    2023-2026 "超高速データサイエンス"

Academic Service

  1. Reviewer

    ICCV 2023, CVPR 2023, IROS 2023-2024, ICRA 2024, TPAMI, WACV 2022, ICPR 2022, 3DV 2020-2021, MVA 2021