1b-16b Variable Bit Precision DNN Processor for Emotional HRI System in Mobile Devices

Chang Hyeon Kim; Jin Mook Lee; Sang Hoon Kang; Sang Yeob Kim; Dong Seok Im; Hoi Jun Yoo

doi:10.23075/jicas.2020.6.3.001

Chang Hyeon Kim Korea Advanced Institute of Science and Technology
Jin Mook Lee Korea Advanced Institute of Science and Technology
Sang Hoon Kang Korea Advanced Institute of Science and Technology
Sang Yeob Kim Korea Advanced Institute of Science and Technology
Dong Seok Im Korea Advanced Institute of Science and Technology
Hoi Jun Yoo Korea Advanced Institute of Science and Technology

DOI: https://doi.org/10.23075/jicas.2020.6.3.001

Keywords: Deep Learning, Deep Learning ASIC, Deep Neural Network, Emotion Recognition, Mobile Deep Learning

Abstract

We propose an energy-efficient DNN processor with the proposed look-up-table-based processing engine (LPE) and near-zero skipper. A CNN-based facial emotion recognition model and an RNN-based emotional dialogue generation model are integrated for the natural human-robot interaction (HRI) system, and it is evaluated by the proposed processor. LPE supports 1 to 16 bit variable weight bit precision, and it achieves 57.6% and 28.5% lower energy consumption than the conventional multiplier-accumulator (MAC) units in 1-16 bit weight precision. Furthermore, the near-zero skipper reduces 36% of MAC operations and consumes 28% lower energy consumption in facial emotion recognition tasks. Implemented in 65 nm CMOS process, the proposed processor occupies 1784×1784 μm² areas and dissipates 0.28 mW and 34.4 mW at 1 frame-per-second (fps) and 30 fps facial emotion recognition tasks.

Author Biographies

Chang Hyeon Kim, Korea Advanced Institute of Science and Technology

Chang Hyeon Kim (S’16) received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology, Daejeon, South Korea, in 2014 and 2016, respectively, where he is currently pursuing the Ph.D. degree. His current research interests include low-power system-on-chip design, especially focused on parallel processor for artificial intelligence and machine learning algorithms.

Jin Mook Lee, Korea Advanced Institute of Science and Technology

Jin Mook Lee (S’15) received the B.S. degrees in electrical engineering from Hanyang University, Seoul, South Korea, in 2014, and the M.S. degree in electrical engineering from the Korea Advanced Institute of Science and Technology, Daejeon, South Korea, in 2016, where he is currently pursuing the Ph.D. degree. His current research interests include energy-efficient deep learning inference/training accelerator ASIC design, embedded deep learning platform design and verification, embedded system development with FPGA programming, and deep learning algorithm for sequence recognition.

Sang Hoon Kang, Korea Advanced Institute of Science and Technology

Sang Hoon Kang (S’16) received the B.S. and M.S. degrees with the School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea, in 2016. He is currently pursuing the Ph.D. degree with the School of Electrical Engineering. His current research interests include low-power vision system-on-chip design and deep learning processor design.

Sang Yeob Kim, Korea Advanced Institute of Science and Technology

Sang Yeob Kim (S’18) received the B.S. degree in electrical engineering from the Korea Advanced Institute of Science and Technology, Daejeon, South Korea, in 2018, where he is currently pursuing the M.S. degree. His current research interests include low-power system-on-chip design, deep neural network accelerators, and machine learning algorithms for deep learning.

Dong Seok Im, Korea Advanced Institute of Science and Technology

Dong Seok Im (S’18) received the B.S. degree in electrical engineering from the Pohang University of Science and Technology (POSTECH), Pohang, South Korea, in 2018. Currently, he is working toward the M.S. degree in electrical engineering at the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea. His current research interests include energy-efficient deep learning SoC design & intelligent vision system.

Hoi Jun Yoo, Korea Advanced Institute of Science and Technology

Hoi Jun Yoo graduated from the Electronic Department, Seoul National University, Seoul, South Korea, in 1983, and the M.S. and Ph.D. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea, in 1985 and 1988, respectively.

He has served as a member for the Executive Committee of ISSCC, Symposium on VLSI, and A-SSCC, the TPC Chair for the A-SSCC 2008 and ISWC 2010, the IEEE Distinguished Lecturer from 2010 to 2011, the Far East Chair for the ISSCC from 2011 to 2012, the Technology Direction Sub-Committee Chair for the ISSCC in 2013, the TPC Vice Chair for the ISSCC in 2014, and the TPC Chair for the ISSCC in 2015.

Homepage : https://ssl.kaist.ac.kr/