Reproduce “Learning to Live: A Step Toward Lifelong Embodied Intelligence” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lirong Che, Peidong Wang, Yimeng Su, Jiong Wang, Yang Chen, Wenbo Fu, Zhenyu Huang, Tiancheng Xu and 6 more
LIFE links observation, selective memory, and reusable tools. RoboBeats gates incoming observations, while Plastic Memory retains selected images and structured records for later questions; EDM reuses a capability-development process from ReSight. On SSS-Bench, RoboBeats-2B improves accuracy by 5.2 percentage points over uniform sampling for observations lasting at least 20 minutes under a matched writing budget; this does not establish lifelong autonomy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.