Project website for Benchmarking Risk-Aware Task-Level Decision Making for Robotic Manipulation with Large Multimodal Models.
Website · Code · Paper and appendix
Lei Zhang, Ju Dong, Kaixin Bai, Minheng Ni, Yu Fu, Zoltán-Csaba Márton, Zhaopeng Chen, and Jianwei Zhang.
The website presents the 23-task benchmark, research findings, 23 simulation clips, and Franka and UR5e demonstrations. All figures, videos, fonts, and PDFs are included locally.
python3 -m http.server 8000Open http://localhost:8000. No package installation or build step is required.
- Edit the page content in
index.html. Its CSS and JavaScript are bundled inline;style.cssandsite.jsretain the matching source copies. Keep those copies consistent when changing styles or interactions. - Store images and videos under
assets/, and published PDFs underdownloads/. downloads/manuscript.pdfis the full paper with the complete appendix. The page uses one combined paper-and-appendix download.- Keep the author order and affiliations consistent with the paper. The citation is available in
citation.biband on the page. - Push changes to
main. GitHub Pages publishes the repository root;.nojekyllkeeps this static site unchanged during deployment.
This repository contains the public project website. Benchmark implementation and usage instructions are maintained in the code repository.