Bookmark Saya

Semantic Scholar Open Access 2019 38 sitasi

A 55-nm, 1.0–0.4V, 1.25-pJ/MAC Time-Domain Mixed-Signal Neuromorphic Accelerator With Stochastic Synapses for Reinforcement Learning in Autonomous Mobile Robots

Anvesha Amaravati Saad Bin Nasir Justin Ting Insik Yoon A. Raychowdhury
Lihat Sumber DOI

Abstrak

Reinforcement learning (RL) is a bio-mimetic learning approach, where agents can learn about an environment by performing specific tasks without any human supervision. RL is inspired by behavioral psychology, where agents take actions to maximize a cumulative reward. In this paper, we present an RL neuromorphic accelerator capable of performing obstacle avoidance in a mobile robot at the edge of the cloud. We propose an energy-efficient time-domain mixed-signal (TD-MS) computational framework. In TD-MS computation, we demonstrate that the energy to compute is proportional to the importance of the computation. We leverage the unique properties of stochastic networks and recent advances in Q-learning in the proposed RL implementation. The 55-nm test chip implements RL using a three-layered fully connected neural network and consumes a peak power of 690 $\mu \text{W}$ .

Topik & Kata Kunci

Computer Science

Penulis (5)

A

Anvesha Amaravati

S

Saad Bin Nasir

J

Justin Ting

I

Insik Yoon

A

A. Raychowdhury

Format Sitasi

APA MLA BibTeX
Amaravati, A., Nasir, S.B., Ting, J., Yoon, I., Raychowdhury, A. (2019). A 55-nm, 1.0–0.4V, 1.25-pJ/MAC Time-Domain Mixed-Signal Neuromorphic Accelerator With Stochastic Synapses for Reinforcement Learning in Autonomous Mobile Robots. https://doi.org/10.1109/JSSC.2018.2881288

Akses Cepat

PDF tidak tersedia langsung

Cek di sumber asli →
Lihat di Sumber doi.org/10.1109/JSSC.2018.2881288
Informasi Jurnal
Tahun Terbit
2019
Bahasa
en
Total Sitasi
38×
Sumber Database
Semantic Scholar
DOI
10.1109/JSSC.2018.2881288
Akses
Open Access ✓