没有合适的资源?快使用搜索试试~ 我知道了~
首页AI accelerator design.pdf
资源详情
资源评论
资源推荐

1
AI Accelerator design & lab
-Hardware IP (SoC) & Memory Design for
Deep Neural Network and Its Application
Date : 2019.04.08
Author : Hyun Kim
Affiliation : Seoul National University of Science and Technology
Electrical and Information Engineering
Position : Assistant Professor
Contact : hyunkim@seoultech.ac.kr / 010-9600-5427
한양대 IDEC 교육

2
Contents
Self-Introduction
HW IP / SoC Design of Deep Neural Networks Platform in Autonomous Driving
HW IP / SoC Design of Memory for Deep Neural Networks
Conclusion
Introduction
Additional Research Topics related to Memory Design

3
Contents
Self-Introduction
HW IP / SoC Design of Deep Neural Networks Platform in Autonomous Driving
HW IP / SoC Design of Memory for Deep Neural Networks
Conclusion
Introduction
Additional Research Topics related to Memory Design

4
About Myself
Self-Introduction
Employment Background
Duration Affiliation Department Title
2018.09~Current
Seoul National University of
Science and Technology
Electrical and Information Engineering Assistant Professor
2016.03~2018.08 Seoul National University BK21+ / Electrical and Computer Engineering BK Professor
2015.03~2016.02 Seoul National University BK21+ / Electrical and Computer Engineering Post Doctor
Education Background
Course Institution Department Duration
Bachelor’s Seoul National University Electrical and Computer Engineering 2005.03~2009.02
Master’s Seoul National University Electrical and Computer Engineering 2009.03~2011.02
Doctorate Seoul National University Electrical and Computer Engineering 2011.03~2015.02
Title of
Dissertation
M.S.
Power-Aware Design with Various Low-Power Algorithms for an
H.264/AVC Encoder
Advisor
M.S. Hyuk-Jae Lee
Ph.D.
H.264-based Low Power Heterogeneous Video Recording System
Ph.D. Hyuk-Jae Lee
Personal Information
Name Date of Birth Nationality
Hyun Kim 1987-02-11 Republic of Korea

5
Research Experience
Video
Compression
System
Optimization
Digital
System (SoC)
Design
Low-Power
Design
Deep
Learning
Resource-Performance
Trade-off Design
R-D Performance vs.
complexity optimization
for Video Codec
Power Consumption vs.
Accuracy optimization
for Deep Learning
Memory architecture &
access optimization
Video Coding Standard
H.264 algorithm & HW design
HEVC algorithm & HW design
Embedded Compression Design
DWT+SPIHT algorithm & HW design
Multi-Codec Design
Codec Integration for best performance
Low-complexity CNN
GPU/RTL based low-complexity CNN
CNN for autonomous driving
Memory for CNN
Approximate DRAM Architecture
Next generation memory: HBM/PCRAM
Processor-In-Memory
HW IP/SoC
Design for
Multimedia
System
Self-Introduction
Multimedia processing
for low-power
Low-power Video Codec
Low-power Deep
Learning
Pre-processing scheme
for low-complexity
Multimedia
HW acceleration
Optimization
Front-end SoC design
HW acceleration
Post-layout simulation
Front-end (FPGA) verification
HW/SW co-design
剩余112页未读,继续阅读

















安全验证
文档复制为VIP权益,开通VIP直接复制

评论0