Reproduce “LASER: Learning Active Sensing for Continuum Field Reconstruction” 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
Huayu Deng, Jinghui Zhong, Xiangming Zhu, Yunbo Wang, Xiaokang Yang
LASER actively repositions sensors to reconstruct evolving physical fields from sparse observations. It encodes measurements into latent states, predicts future states, and uses a policy to choose sensor movements, with reconstruction error as reward. On a Navier–Stokes test with 64 sensors, its MSE was 0.169 versus 26.61 for the best heuristic, on a ×10⁻³ scale; results are simulations, not physical deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.