홍상현 오레곤 주립 대학교

Oregon State University
Computer Science Dept.
Cybersecurity | AI
Contact Information

Office: Room 4103, Kelley Engineering Center (KEC)
2500 NW Monroe Ave
Corvallis, OR 97331 USA
Office Hours: Tu/Th: 2 - 3 pm

          

Press

04.2022

TechXplore
Techradar.pro

06.2021

TechTalks

05.2021

Dev Podcast
MIT Tech Review

02.2021

USENIX Enigma 2021
(Ted Talk for Security)

Teaching

Spring 23 CS370: Intro to Sec.
CS499/599: MLSec.
Winter 22 CS344: OS I
Spring 22 CS344: OS I
Winter 21 CS499/599: MLSec.
Students [Full list]

Hoang Le (PhD, CS)
Gabriel Ritter (PhD, CS,
  co-advise w. Rakesh Bobba)
Eunjin Roh (MS, CS)

Alumni

2022: Ryan Little (BS, CS)
  Now a PhD student at UMD

I am an Assistant Professor of Computer Science at Oregon State University.
I work at the intersection of computer security, privacy and machine learning.

Research Interests

My research objective is to build trustworthy machine learning (ML) systems—systems that employ ML models as a key component—so that humans use those AI-enabled systems to improve our lives and society in the future. Thus far, I’ve been interested in characterizing the security/privacy and dependability issues of ML systems from a systems security perspective. I am selected as a DARPA Riser (2022) and was invited as a speaker at USENIX Enigma (2021).

Please drop me an email with your CV if you're interested in working with me.

Bio

I earned my Ph.D. from the University of Maryland, College Park, under the supervision of Dr. Tudor Dumitras in 2021. I received my bachelor's degree from Seoul National University in 2015. I was fortunate to spend a winter at Google Brain in 2021 (working with Nicholas Carlini and Alexey Kurakin) and to spend 6-months at Frame.io in 2017 (working with Abhinav Srivastava).

News


Jul. 16, 2022
One paper is accepted at IEEE VIS 2022
Jun. 13, 2022
One paper is accepted at ICML Workshop on Continuous-Time Methods for ML 2022
May. 10, 2022
Selected as a DARPA Riser 2022.
Mar. 31, 2022
My student (Hoang)'s first paper is on arXiv; congratulations!
Apr. 8, 2022
One paper is accepted at ISSTA 2022.
Jan. 22, 2022
One paper is accepted at ICLR 2022.
Sep. 28, 2021
One paper is accepted at NeurIPs 2021.
Sep. 16, 2021
Joined CS @ Oregon State University as an Assistant Professor.

Selected Publications [Full list]


Truth Serum: Poisoning Machine Learning Models to Reveal Their Secrets
Florian Tramèr, Reza Shokri, Ayrton San Joaquin, Hoang Le, Matthew Jagielski, Sanghyun Hong,
Nicholas Carlini
(*authors ordered reverse-alphabetically)
arXiv Preprint 2022.
PDF | Code | Media

Improving Cross-Platform Binary Analysis Using Representation Learning via Graph Alignment
Geunwoo Kim, Sanghyun Hong, Michael Franz, Dokyung Song
The ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) 2022. To Appear
PDF | Code

Data Poisoning Won't Save You From Facial Recognition
Evani Radiya-Dixit, Sanghyun Hong, Nicholas Carlini, Florian Tramer
International Conference on Learning Representations (ICLR) 2022.
PDF | Code | Poster

Qu-ANTI-zation: Exploiting Neural Network Quantization for Achieving Adversarial Outcomes
Sanghyun Hong, Michael-Andrei Panaitescu-Liess, Yigitcan Kaya, and Tudor Dumitraș
Advances in Neural Information Processing Systems (NeurIPS) 2021.
PDF | Code | Poster

A Panda? No, It's a Sloth: Slowdown Attacks on Adaptive Multi-Exit Neural Network Inference
*Sanghyun Hong, *Yigitcan Kaya, Ionuţ-Vlad Modoranu, and Tudor Dumitraș(* equal contribution)
International Conference on Learning Representations (ICLR) 2021. [Spotlight]
PDF | Code | Spotlight Presentation

How to 0wn NAS in Your Spare Time
Sanghyun Hong, Michael Davinroy, Yigitcan Kaya, Dana Dachman-Soled, and Tudor Dumitraș
International Conference on Learning Representations (ICLR) 2020.
PDF | Code | Poster

Terminal Brain Damage: Exposing the Graceless Degradation in Deep Neural Networks
Under Hardware Fault Attacks

Sanghyun Hong, Pietro Frigo, Yigitcan Kaya, Cristiano Giuffrida, and Tudor Dumitraș
Proceedings of The 28th USENIX Security Symposium (USENIX Security) 2019.
PDF | Presentation

Shallow-Deep Networks: Understanding and Mitigating Network Overthinking
Yigitcan Kaya, Sanghyun Hong, and Tudor Dumitraș
International Conference on Machine Learning (ICML) 2019.
PDF | Code

Go Serverless: Securing Cloud via Serverless Design Patterns
Sanghyun Hong, Abhinav Srivastava, William Shambrook, and Tudor Dumitraș
10th USENIX Workshop on Hot Topics in Cloud Computing (HotCloud) 2018.
PDF | Slides