Reconstructing Dynamics, Not Values: A Delay-Embedding Approach to Gappy Time Series
bravo_abad · x · 2026-09-20
Instead of predicting missing points from correlations, Wu et al. recover the geometry of the dynamical system first. Delay embedding reconstructs the attractor from a fully observed variable; a Gaussian process then learns the mapping between the partial and full manifolds to fill missing states. The key idea: dynamical-systems theory supplies the prior, so ML doesn't rediscover the dynamics from scratch.
More from Research
- MIT Media Lab hosts ScienceClaw Hackathon to build 'internet of agents' for science — ProfBuehlerMIT · 2026-09-20
- 83k PTB trace downloads in a month: researcher warns RL agents may be reward hacking via HF lookups — maksym_andr · 2026-09-20
- Schmidhuber revisits history: 1971 deep network modeled the British economy — SchmidhuberAI · 2026-09-20
- Qwen researcher: human annotation may be noisier than auto-labeling — antoine_chaffin · 2026-09-20
- Retrieval eval datasets are error-prone; LLM relabeling proposed as fix — CShorten30 · 2026-09-20
- PhyFilter from Beihang and NTU lets sim-trained quadrupeds walk real terrain, in npj Robotics — jiqizhixin · 2026-09-20