Preliminary Results Towards Reinforcement Learning with Mixed-Signal Memristive Neuromorphic Circuits

Nan Wu (1) , Adrien F Vincent (2, 1) , Dmitri Strukov (1)

1 ECE-UCSB - Department of Electrical and Computer Engineering [Santa Barbara]
2 IMS - Laboratoire de l'intégration, du matériau au système

Mots clés: Artificial neural networks; Reinforcement learning; Mem-ristor; ReRAM; In-situ training; Hardware implementation; Actor-Critic mode

Domaines: Micro et nanotechnologies/Microélectronique; Electronique; Sciences de l'ingénieur [physics]

Nan Wu, Adrien F Vincent, Dmitri Strukov. Preliminary Results Towards Reinforcement Learning with Mixed-Signal Memristive Neuromorphic Circuits. 2019 IEEE International Symposium on Circuits and Systems (ISCAS), May 2019, Sapporo, Japan. pp.1-5, ⟨10.1109/ISCAS.2019.8702229⟩. ⟨hal-02487827⟩