Erlangen 2022 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
Q: Fachverband Quantenoptik und Photonik
Q 15: Quantum Information (Quantum Computing and Simulation)
Q 15.1: Hauptvortrag
Dienstag, 15. März 2022, 10:30–11:00, Q-H12
A hybrid quantum classical learning agent — •Sabine Wölk — Institute of Quantum Technologies, German Aerospace Center (DLR), Ulm, Germany
Machine learning and quantum information become more and more important in our digital world. An important paradigm within machine learning is reinforcement learning. Here, a decision-making entity called agent solves a task by interacting with its environment. The agent updates its behaviour, and thus learns, by using the obtained feedback it receives from the environment. We can speed up the learning if the agent and its environment can be transformed into corresponding quantum systems interacting with each other.
We have developed a hybrid quantum classical learning agent which combines quantum exploration of the environment with classical behavior updates [1,2]. In this way, we can achieve a quadratic speedup in learning. In this talk, I will explain the main features of this hybrid learning agent and discuss possible applications as well as first proof-of-principle experiments.
[1] A. Hamann and S. Wölk, Performance analysis of a hybrid agent for quantum-accessible reinforcement learning, arXiv: 2107.14001.
[2] V. Saggio, B. Asenbeck, A. Hamann, T. Strömberg, P. Schiansky, V. Dunjko, N. Friis, N. C. Harris, M. Hochberg, D. Englund, S. Wölk, H. J. Briegel, and P. Walther, Experimental quantum speed-up in reinforcement learning agents, Nature 591, 229 (2021).