About Me
I’m CSE student/Ph.D candidate at CMALab, Seoul National University.
My general topic of interest is hardware-software co-design.
Specifically, from application down to hardware:
Publication
NIPQ: Noise proxy-based Integrated Pseudo-Quantization
I was accidentally not included in the author list. You can find me in the Acknowledgements section.
CVPR 2023
Juncheol Shin, Junhyuk So, Sein Park, Seungyeop Kang, Sungjoo Yoo, and Eunhyeok Park
Differentiable mixed precision QAT, part of IITP project.
BlindFilter: Privacy-Preserving Spam Email Detection Using Homomorphic Encryption
SRDS 2023
Dongwon Lee, Myeonghwan Ahn, Hyesun Kwak, Jin B. Hong and Hyoungshick Kim
Pseudo quantization noise as a defensive measure against adversarial attack on parameters of homomorphic encryption algorithm.
Projects
Lab cluster networking and management
- Designed and implemented networking for cluster in our lab.
- Network consists of two leaf-spine fabrics: one for management and the other for data.
- Each fabric is implemented as OSPF + ECMP with L3 HW offloading.
- Implemented all-flash Ceph storage cluster.
- Implemented management system for the whole cluster using canonical MAAS.
IITP governmental project
Development of model compression framework for scalable on-device AI computing on Edge applications. MSIT, No.2019-0-01906, No.2021-0-00105, and No.2021-0-00310
I’m contributing to the project starting from 2H 2021. My focus is devising a differentiable mixed-precision quantization algorithm and proving its effectiveness on various applications.
From SNUCSE 2020 “creative integrated design” class.
Trash classification on Arduino Nano 33 BLE and Raspberry Pi 2.
From training various models(MbV1, MbV2, ShV2) on resized imagenet(96x96, 128x128, …), fine-tuning on trashnet down to actually running on hardwares.
Coursework TA
Hardware System Design
2022, 2023
This course covers hardware-software co-design in terms of building custom NN accelerator.
Till 2022, the course covered: Xilinx FPGA (Zedboard) ~ Vivado ~ verilog ~ C++ ~ MNIST classification(caffe2).
In 2023, I revised the course to be: Amaranth simulator ~ simple NN compiler ~ torch.
- GEMM accelerator
- simple instruction set
- reconfigurable
- outer product matrix-matrix multiplication
- on Amaranth
- simple NN compiler
- generate simple instruction set code given input pytorch GEMM
This course covers latest topics in ML applications overall.
In 2022, I covered YOLOv5 and applying NIPQ on it.
Education
Graduate
2021.9 ~
Ph.D candidate at CMALab, Seoul National University
Undergraduate
2015.3 ~ 2017.7 // 2019.7 ~ 2021.8
Seoul National University, Computer Science & Engineering
Experience
CMALab, Seoul National University
Internship
2020.12 ~ 2021.9
My undergraduate thesis was on applying differentiable NAS(with Gumbel-softmax), choosing [Binary | INT8] layerwise, targeting image classification network on TinyML.
It was extension of lightb0x/arduino_trash_classification
Samsung Electronics
Internship
2020.1.6 ~ 2020.2.14
In cloud management group, objective of the team was to try out new technologies, especially Kubernetes and Terraform at that time.
I tried toy project with React, golang-gin, Docker, Kubernetes, Terraform.