@article{2022_3C_ppg,
  title={Recurrent Neural Network Controllers Synthesis with Stability Guarantees for Partially Observed Systems},
  author={Gu, Fangda and Yin, He and El Ghaoui, Laurent and Arcak, Murat and Seiler, Peter and Jin, Ming},
  year={2022},
   journal = "AAAI Conference on Artificial Intelligence (AAAI)",
  url_pdf={PPG_2021.pdf},
  url_arXiv={https://arxiv.org/abs/2109.03861},
  keywords = {Machine learning, Reinforcement learning},
  abstract={Neural network controllers have become popular in control tasks thanks to their flexibility and expressivity. Stability is a crucial property for safety-critical dynamical systems, while stabilization of partially observed systems, in many cases, requires controllers to retain and process long-term memories of the past. We consider the important class of recurrent neural networks (RNN) as dynamic controllers for nonlinear uncertain partially-observed systems, and derive convex stability conditions based on integral quadratic constraints, S-lemma and sequential convexification. To ensure stability during the learning and control process, we propose a projected policy gradient method that iteratively enforces the stability conditions in the reparametrized space taking advantage of mild additional information on system dynamics. Numerical experiments show that our method learns stabilizing controllers while using fewer samples and achieving higher final performance compared with policy gradient.}
}

{"_id":"x4aqNKGubN9YJiTRj","bibbaseid":"gu-yin-elghaoui-arcak-seiler-jin-recurrentneuralnetworkcontrollerssynthesiswithstabilityguaranteesforpartiallyobservedsystems-2022","author_short":["Gu, F.","Yin, H.","El Ghaoui, L.","Arcak, M.","Seiler, P.","Jin, M."],"bibdata":{"bibtype":"article","type":"article","title":"Recurrent Neural Network Controllers Synthesis with Stability Guarantees for Partially Observed Systems","author":[{"propositions":[],"lastnames":["Gu"],"firstnames":["Fangda"],"suffixes":[]},{"propositions":[],"lastnames":["Yin"],"firstnames":["He"],"suffixes":[]},{"propositions":[],"lastnames":["El","Ghaoui"],"firstnames":["Laurent"],"suffixes":[]},{"propositions":[],"lastnames":["Arcak"],"firstnames":["Murat"],"suffixes":[]},{"propositions":[],"lastnames":["Seiler"],"firstnames":["Peter"],"suffixes":[]},{"propositions":[],"lastnames":["Jin"],"firstnames":["Ming"],"suffixes":[]}],"year":"2022","journal":"AAAI Conference on Artificial Intelligence (AAAI)","url_pdf":"PPG_2021.pdf","url_arxiv":"https://arxiv.org/abs/2109.03861","keywords":"Machine learning, Reinforcement learning","abstract":"Neural network controllers have become popular in control tasks thanks to their flexibility and expressivity. Stability is a crucial property for safety-critical dynamical systems, while stabilization of partially observed systems, in many cases, requires controllers to retain and process long-term memories of the past. We consider the important class of recurrent neural networks (RNN) as dynamic controllers for nonlinear uncertain partially-observed systems, and derive convex stability conditions based on integral quadratic constraints, S-lemma and sequential convexification. To ensure stability during the learning and control process, we propose a projected policy gradient method that iteratively enforces the stability conditions in the reparametrized space taking advantage of mild additional information on system dynamics. Numerical experiments show that our method learns stabilizing controllers while using fewer samples and achieving higher final performance compared with policy gradient.","bibtex":"@article{2022_3C_ppg,\n title={Recurrent Neural Network Controllers Synthesis with Stability Guarantees for Partially Observed Systems},\n author={Gu, Fangda and Yin, He and El Ghaoui, Laurent and Arcak, Murat and Seiler, Peter and Jin, Ming},\n year={2022},\n journal = \"AAAI Conference on Artificial Intelligence (AAAI)\",\n url_pdf={PPG_2021.pdf},\n url_arXiv={https://arxiv.org/abs/2109.03861},\n keywords = {Machine learning, Reinforcement learning},\n abstract={Neural network controllers have become popular in control tasks thanks to their flexibility and expressivity. Stability is a crucial property for safety-critical dynamical systems, while stabilization of partially observed systems, in many cases, requires controllers to retain and process long-term memories of the past. We consider the important class of recurrent neural networks (RNN) as dynamic controllers for nonlinear uncertain partially-observed systems, and derive convex stability conditions based on integral quadratic constraints, S-lemma and sequential convexification. To ensure stability during the learning and control process, we propose a projected policy gradient method that iteratively enforces the stability conditions in the reparametrized space taking advantage of mild additional information on system dynamics. Numerical experiments show that our method learns stabilizing controllers while using fewer samples and achieving higher final performance compared with policy gradient.}\n}\n\n","author_short":["Gu, F.","Yin, H.","El Ghaoui, L.","Arcak, M.","Seiler, P.","Jin, M."],"key":"2022_3C_ppg","id":"2022_3C_ppg","bibbaseid":"gu-yin-elghaoui-arcak-seiler-jin-recurrentneuralnetworkcontrollerssynthesiswithstabilityguaranteesforpartiallyobservedsystems-2022","role":"author","urls":{" pdf":"http://www.jinming.tech/papers/PPG_2021.pdf"," arxiv":"https://arxiv.org/abs/2109.03861"},"keyword":["Machine learning","Reinforcement learning"],"metadata":{"authorlinks":{}},"downloads":6},"bibtype":"article","biburl":"http://www.jinming.tech/papers/myref.bib","dataSources":["sTzDHHaipTZWjp8oe","Y64tp2HnDCfXgLdc5"],"keywords":["machine learning","reinforcement learning"],"search_terms":["recurrent","neural","network","controllers","synthesis","stability","guarantees","partially","observed","systems","gu","yin","el ghaoui","arcak","seiler","jin"],"title":"Recurrent Neural Network Controllers Synthesis with Stability Guarantees for Partially Observed Systems","year":2022,"downloads":6}