Reinforcement Learning for Athletic Intelligence: Lessons from the 1st “AI Olympics with RealAIGym” Competition
Paper i proceeding, 2024

As artificial intelligence gains new capabilities, it becomes important to evaluate it on real-world tasks. In particular, the fields of robotics and reinforcement learning (RL) are lacking in standardized benchmarking tasks on real hardware. To facilitate reproducibility and stimulate algorithmic advancements, we held an AI Olympics competition at IJCAI 2023 conference based on the double pendulum system in the RealAIGym project where the participants were asked to develop a controller for the swing up and stabilization task. This paper presents the methods and results from the top participating teams and provides insights into the real-world performance of RL algorithms with respect to a baseline time-varying LQR controller.

Författare

Felix Wiebe

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Niccolò Turcato

Università di Padova

Alberto Dalla Libera

Università di Padova

Chi Zhang

Technische Universität München

Theo Vincent

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Shubham Vyas

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Giulio Giacomuzzo

Università di Padova

Ruggero Carli

Università di Padova

Diego Romeres

Mitsubishi Electric Research Laboratories

Akhil Sathuluri

Technische Universität München

Markus Zimmermann

Technische Universität München

Boris Belousov

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Jan Peters

Technische Universität Darmstadt

Centre for Cognitive Science

Hessian.AI

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Frank Kirchner

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Universität Bremen

Shivesh Kumar

Deutsches Forschungszentrum fur Kunstliche Intelligenz

Chalmers, Mekanik och maritima vetenskaper, Dynamik

Forskning Andra publikationer

IJCAI International Joint Conference on Artificial Intelligence

10450823 (ISSN)

8833-8837
9781956792041 (ISBN)

33rd International Joint Conference on Artificial Intelligence, IJCAI 2024
Jeju, South Korea,

Ämneskategorier

Robotteknik och automation

Reglerteknik

Datavetenskap (datalogi)

Mer information

Senast uppdaterat

2024-09-27

Feedback och support

Om du har frågor, behöver hjälp, hittar en bugg eller vill ge feedback kan du göra det här nedan. Du når oss också direkt per e-post research.lib@chalmers.se.

Om tjänsten

Research.chalmers.se innehåller information om forskning på Chalmers, publikationer och projekt inklusive information om finansiärer och samarbetspartners.

Läs mer om tjänsten, täckningsgrad och vilka som kan se informationen

Personuppgifter och cookies

Tillgänglighet

Citation Style Language
citeproc-js (Frank Bennett)

Länkar

Chalmers bibliotek
Chalmers forskning
Chalmers examensarbeten

Chalmers tekniska högskola

412 96 GÖTEBORG
TELEFON: 031-772 10 00
WWW.CHALMERS.SE