Embedded Files
About Me
About Me
Hi there, thanks for dropping by my web. You know my name already :). I am a 3rd year Ph.D. student in Computer Engineering at University of Southern California working under supervision of Prof. Massoud Pedram (SPORT Lab) and Prof. Peter A. Beerel (HAL@usc) . I was born in India and did my B.Tech and M.Tech in Electronics Engineering from West Bengal University of Technology and Indian Institute of Technology, Kharagpur, respectively. My technical interest includes hardware friendly Computer vision, hardware accelerator architecture design for neural networks, In/near memory computing, computer architecture. Apart from all these catchy words I love long drive along pacific highway, traveling, playing and watching soccer.
To have a brief information about my experience and academic background you can view my resume
Google scholar: https://scholar.google.com/citations?user=b591SVkAAAAJ&hl=en
Contact Information
Contact Information
Ming Hsieh Department of Electrical Engineering
University of Southern California
3740 McClintock Ave, EEB 314
Los Angeles CA 90089-2562
email: souvikku@usc.edu
Linkedin: https://www.linkedin.com/in/souvik-kundu-64922b50/
Selected Projects
Selected Projects
Are dense CNN kernels really necessary?
Are dense CNN kernels really necessary?
A study of sparsely represented kernels, where we propose hardware friendly sparse models enabling low cost data-transfer from DRAM (IEEE TC 2020, Allerton 2019, Best poster at USC research festival 2019)
Making MCA based cross-bar array more compact
Making MCA based cross-bar array more compact
We propose a new form of pre-defined sparsity that can reduce the cross bar array size drastically, thus helps in making RRAM based ML accelerators scalable (ISVLSI 2019)
A novel ALU architecture for SFQ-technology
A novel ALU architecture for SFQ-technology
We propose a novel block-skewing technique to increased the throughput and reduce initiation interval of gate-level pipelined SFQ-logic based ALUs (IEEE ISEC 2019)
Publications
Publications
2020
- S. Kundu, M. Nazemi, M. Pedram, K. M. Chugg, P. A. Beerel, "Pre-defined Sparsity for Low-Complexity Convolutional Neural Networks", accepted in IEEE Transactions on Computers, 2020 (Journal)
2019
- S. Kundu, S. Prakash, H. Akrami, P. A. Beerel, K. M. Chugg, "pSConv: A Pre-defined Sparse Kernel Based Convolution for Deep CNNs", Allerton Conference 2019.
- A. Fayyazi *, S. Kundu *, S. Nazarian, P.A. Beerel, M. Pedram , "CSrram: Area-Efficient Low-Power Ex-Situ Training Framework for Memristive Neuromorphic Circuits Based on Clustered Sparsity ", ISVLSI 2019.
- S. Kundu, G. Datta, M. Pedram, P. A. Beerel, "qBSA: Logic Design of a 32-bit Block-Skewed RSFQ Arithmetic Logic Unit", ISEC 2019.
- G. Datta, H. Cong, S. Kundu, P. A. Beerel, "qCDC: Metastability-Resilient Synchronization FIFO for SFQ Logic", ISEC 2019.
2018
- S. Dey, D. Chen, Z. Li, S. Kundu, K.W. Huang, K. Chugg and P.A. Beerel, "A Highly Parallel FPGA Implementation of Sparse Neural Network Training", ReConFig 2018.
(* = equal contribution)
Industry Work Experiences
Industry Work Experiences
Deep Learning Research Intern, Intel AI Labs, Santa Clara, USA. May 2020 to Aug 2020
Digital Design Engineer, Texas Instruments, India. (Duration - June 2016 to July 2017)
R & D Engineer II, Synopsys, India. (Duration - June 2015 to June 2016)
Services
Services
Reviewer at:
- BMVC 2020
- Micro 2020
- DAC 2020 (external)
- ISCAS 2020
- IEEE Transactions on Neural Networks and Learning Systems
- IEEE Transactions on Computers
- IEEE Transactions on Circuits and Systems I
- IEEE Transactions on Circuits and Systems II
Student PC member at:
- EMNLP 2020
Mentored undergraduate students in:
- IUSSTF: Viterbi India internship program (2018, 2020)
Google Sites
Report abuse