Reproduce “CERNet: Class-Embedding Predictive-Coding RNN for Unified Robot Motion, Recognition, and Confidence Estimation” 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

Hiroki Sawada, Alexandre Pitti, Mathias Quoy

CERNet uses one predictive-coding model to generate movements and infer their class from observed trajectories. A class vector conditions generation and is adjusted online during inference as prediction errors update internal states. In simulation, MultiLarge achieved DTW 0.25 ± 0.11 versus 1.03 ± 0.70 for SingleLarge, a reported 76% lower error. Tests covered 26 taught alphabet trajectories on a robot arm.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.