Changhoon Kim
@ChanghoonKim3
Followers
163
Following
502
Media
6
Statuses
62
Assistant Professor at Soongsil University Ex-AWS; PhD @ASU APG; BJJ Blue Belt
Joined February 2020
Join us as we explore how to build trustworthy, robust, and safe vision-language generative models (e.g., Text-to-Image & Image-to-Text models). #ICCV2025 #GenerativeAI #VisionLanguage #AITrustworthiness #ResponsibleAI
0
0
1
🚀 Excited to announce our ICCV 2025 Tutorial: Responsible Vision-Language Generative Models (RVLGM)! 🗓 October 19 | ICCV 2025, Hawaii 📍 Hawaii Convention 316B 🔗
lnkd.in
This link will take you to a page that’s not on LinkedIn
1
0
2
For Q4 2025, Active Perception Research Group and our collaborators made steady research contributions in video #GenAI, image #GenAI and secure #GenAI. We will present two papers at @NeurIPSConf, another two at @ICCVConference, one at @emnlpmeeting, and one at @TmlrOrg journal.🤠
1
3
7
📍 NAACL 2025 (Mexico City) 🕚 May 2, 11:00–12:30 (local time) 📘 Read the paper: arXiv:2412.03587 Thanks to my amazing co-authors for their collaboration! Hyegang Son, @Yonglak Son, @Younggeun Kim #NAACL2025 #NLP #EfficientTraining #ResearchMilestone #LLM
0
0
1
We introduce SAFE, a method that selectively freezes less useful adapters early in training—cutting memory use by 42.85%, computation amount by 34.59%, and training time by 11.82%, all without sacrificing performance.
1
0
1
Excited to share that my first paper as a corresponding author has been accepted as an 🎤 Oral at NAACL 2025! 📄 "Not All Adapters Matter: Selective Adapter Freezing for Memory-Efficient Fine-Tuning of Language Models"
1
0
2
📢 Come join us for our 1st Workshop on Responsibly Building the Next Generation of Multimodal Foundational Models @NeurIPSConf .
📢 Come join us for our 1st Workshop on Responsibly Building the Next Generation of Multimodal Foundational Models @NeurIPSConf. We have seen great progress in #GenAI and have encountered various challenges in aspects of responsibility. What precautions do we need to take
0
1
7
Working on generative models and attending @eccvconf ? Don’t miss our tutorial: "Responsibly Building Generative Models"! 🗓️ 9/29 (Sunday), 2 PM 📍 Suite 7 (In-person) 💻 Details here: https://t.co/dpuoFezvyH We’ll dive into key #GenAI failures and solutions. See you there! 🙌
asu-apg.github.io
Working on generative models? Attending #ECCV? You should check out our tutorial - Responsibly Building Generative Models. ⏰ - 9/29 (Sunday), 2 PM 📍 - Suite 7 (In-person) 💻 - https://t.co/aqlIDFNJ85 We will discuss different aspects of failures in #GenAI ! #tutorial
0
0
4
Thrilled to announce that our work has been accepted to #NeurIPS2024! Super excited to share our research with the community and be part of such an incredible conference. Thanks a lot, as always. @patelmaitreya @chengshengcs @prof_yz
I am happy to announce that TripletCLIP got accepted at #NeurIPS2024. Congrats to all the authors! Stay tuned for more details. 🎶 @SCAI_ASU @UMBC
1
2
9
My friend @chengshengcs led a new direction using energy-based model for latent space neural ODEs.
Latent Space Energy-based Neural ODEs
0
0
1
🚨 Excited to share that our work, Robust Adversarial Concept Erasure for #T2I (R.A.C.E.), has been accepted for an ORAL presentation at ECCV (@eccvconf)! Please check: https://t.co/ubqWUMf44Y Thank you to @kyleminbs and @prof_yz.
🚨Introducing R.A.C.E.: a novel adversarial training approach to fortify concept erasure without extra modules. 🔹Single-timestep Adv. attack 🔹Defends against white/black-box prompt attacks 🔹Enhances model robustness Read more: https://t.co/ubqWUMfBUw
@kyleminbs, @prof_yz
0
1
11
🚨🚨Excited to announce the 1st Workshop on Responsibly Building the Next Generation of Multimodal Foundation Models @NeurIPSConf 🚨🚨 Paper submission deadline (‼️30th August ‼️) and other details at: https://t.co/MsziIsPlhc
#NeurIPS2024 #WorkshopRBFM24 Key details 👇👇 (🧵)
asu-apg.github.io
2nd Workshop on Responsibly Building the Next Generation of Multimodal Foundational Models.
1
2
12
🚨Introducing R.A.C.E.: a novel adversarial training approach to fortify concept erasure without extra modules. 🔹Single-timestep Adv. attack 🔹Defends against white/black-box prompt attacks 🔹Enhances model robustness Read more: https://t.co/ubqWUMfBUw
@kyleminbs, @prof_yz
0
2
8
I'm proud to announce two papers accepted at CVPR'24: "WOUAF: A Cornerstone for Responsible Generative Models" "ECLIPSE: Advancing Efficiency in Generative Models" Heartfelt thanks to my collaborators Kyle Min, @patelmaitreya , @chengshengcs , and the supportive @prof_yz.
WOUAF:Weight Modulation for User Attribution and Fingerprinting in Text-to-Image Diffusion Models paper page: https://t.co/901aYWWHaH demo: https://t.co/Jkq3vrrdSe The rapid advancement of generative models, facilitating the creation of hyper-realistic images from textual
2
4
17
📣📣We will be presenting ConceptBed today @RealAAAI. If you want to learn more about the robustness and efficiency of T2I models, find me and the team (@prof_yz and @trgokhale) at the poster (#308) session today evening.
🚨Exciting News! “ConceptBed” has been accepted at #AAAI24 main track. I’m thrilled to also present it at the #NeurIPS23 workshop on diffusion models. 🤖💡 Find me and my colleagues at the NeurIPS to discuss more about this work. (details about #NeurIPS23 schedule coming soon)
0
3
9
It was truly an honor to be selected for the Doctoral Consortium at @RealAAAI . Grateful for this incredible opportunity to engage with leading researchers and present my work. @RealAAAI
#DoctoralConsortium
0
0
8
📢 "Frontier Topics in Generative AI" Seminar Series! I am delighted to share this exciting news. At @ASU, @ApgAsu will be hosting a seminar series featuring esteemed researchers. For more information, please refer to the original message.
🚨🚨We are hosting "Frontier Topics in Generative AI" Seminar Series at @ASU . This series delves into the cutting-edge of GenAI, exploring key areas like large-language models, text-to-image, video generation, and more. We have our first speaker this week.
0
0
4
I had the privilege of discussing these insights with Lucy at AIhub. And I extend my heartfelt gratitude to my dear friend Kyle Min and my esteemed advisor 'YZ' Yezhou Yang for their invaluable support and guidance throughout this journey.
0
0
0
I'm pleased to share that my interview with AIhub is now available. This conversation delves into the topic of image generative model reliability and also touches upon my personal experiences during my Ph.D. journey. https://t.co/QQbMwq33PB
aihub.org
1
1
4
It was a fantastic opportunity to present the tutorial on "Reliability of Generative Models in Vision at #WACV2024". Thanks to all who attended and showed interest! Special thanks co-organizers @gowthami_s, @trgokhale, and @prof_yz. #GenerativeAI #ComputerVision
0
0
9