One of our five papers, HEPA, was accepted as a spotlight at ICML. It outperforms SOTA time series foundation models like Chronos and TimesFM at 1/100th the parameter count. The other four papers are linked from our LinkedIn post
https://www.linkedin.com/posts/galombardi_research-ai-roboti... alongside more context on the work. Really excited to get other people's thoughts on the work!