Wrapped up KAIST @ ICML as a student organizer! 🇰🇷✨
Walked away deeply inspired by the research and career insights shared by top global AI scholars.
Huge thanks to all the amazing panelists and attendees who made KAIST @ ICML 2026 so special! 🚀
@KAIST_AI, @icmlconf
🌟 KAIST Graduate School of AI in ICML 2026: Where to Find Our Faculty and Students🔎
KAIST Graduate School of AI will have a strong presence at ICML 2026, with faculty members and students contributing as organizers, keynote speakers, panel organizers, and community event leaders. The events span probabilistic ML, AI for Science, world models, uncertainty in agentic systems, philosophy of machine learning, women in machine learning, and broader KAIST community activities. Full list in chronological order below:
ICML official roles
Chulhee Yun — Social Chair
Cortiq Summit 2026 (7/3)
https://t.co/mqtUoBeBmG
Seong Joon Oh (@coallaoh) — Organizer
ProbML 2026 (7/5)
https://t.co/IhvpMLhMlp
Sungsoo Ahn (@sungsoo_ahn_) —Organizer
Juho Lee — Keynote Speaker
To be held at KAIST Hongneung Campus
KAIST @ ICML 2026 (7/6)
https://t.co/vhrWsHEapc
-
Organizers: Alice Oh (@aliceoh; SoC), Hyunwoo J. Kim (@hyunwoojkim; SoC), Hyunwoo Kim (@hyunw_kim; AI), Jae-Gil Lee (AI Computing), Joyce Jiyoung Whang (SoC), Minhyuk Sung (@MinhyukSung; SoC), Seong Joon Oh (@coallaoh; AI) , SeYoung Yun (AI), Sungjin Ahn (@SungjinAhn_; AI Computing), Uichin Lee (AI Computing)
Students: Euiin Yi (@Yi_Euiin; AI), Eunsu Kim (@euns0o_kim; SoC), Haneul Yoo (@HaneulYoo13; SoC), Jaihoon Kim (@KimJaihoon; SoC), Jihyeok Jung (AI), Jisu Shin (SoC), Junyeong Park (@jjjunyeong; SoC), Minseong Bae (@kylebae1017; SoC), Sunwoo Bae (SoC), Heehyeon Kim (SoC)
AI4Science Korean Gathering @ ICML (7/6)
https://t.co/wcBS3jczJp
Sungsoo Ahn (@sungsoo_ahn_) — Organizer
(EXPO Talk Panel) Seoul World Model: Grounding World Simulation Models in a Real-World Metropolis (7/6)
https://t.co/3ObqPvNLWd
Seungryong Kim (@KimSeungry62571) — Organizer
WiML Symposium @ ICML 2026 (7/8)
https://t.co/hHOcAiLnDe
Hyunjung Shim — Organizer
ICML SPIGM Workshop (7/10)
https://t.co/MwPJADGkXa
Seul Lee (@SeulLee05) — Organizer
ICML AI4Science Workshop (7/11)
https://t.co/bYT5MBHmM8
Sungsoo Ahn (@sungsoo_ahn_) — Organizer
ICML Philosophy Meets Machine Learning Workshop (7/11)
https://t.co/6EGo2eZgKb
Jaesik Choi — Organizer
ICML Statistical Frameworks for Uncertainty in Agentic Systems Workshop (7/11)
https://t.co/xuWrmdkUQT
Seong Joon Oh (@coallaoh) — Keynote Speaker
We look forward to seeing many colleagues at ICML 2026 and related events. Please join us at these events and connect with KAIST AI members throughout ICML week!
🛑 Does post-training teach LLMs new agentic skills, or just sharpen what they already know?
[1/11] Introducing Sharpening Tax 🧾 to measure the hidden cost of post-training!
Multi-domain RLVR trains one model across math, code, logic, science, tables, and simulation. But the key curriculum question is usually fixed or hand-tuned: which domain should we sample next?
🚨 New paper: Transfer-Aware Curriculum (TAC). TAC asks not only “where is the model learning right now?” but also “will this gradient step help the other domains too?”
⚡ The signal is almost free: we reuse GRPO gradients already computed during training, sketch them with random projections, and estimate cross-domain transfer by gradient alignment. No extra rollouts, no held-out probes, <1% wall-clock overhead.
📈 TAC gets the best macro-average accuracy on both Qwen3-1.7B and Llama3.2-3B, with up to +2.8 pts, or 10% relative, over a learnability-only bandit.
😲 Surprise: math was among the least transferable domains, even though RLVR often leans on it most.
Google Student Researcher Program 2026 is now OPEN!
Work on REAL AI/ML projects with:
• Google Research
• DeepMind
• Google Cloud
Open to: Bachelors / Masters / PhD
Duration: 3–12 months
Deadline: March 31
If you're serious about AI, this is your shot.
Apply here
https://t.co/6ZAXdSSczV
@yourgradcoach 100% this. going through the phd interview process and I'm genuinely shocked by how many candidates haven't even skimmed the lab's recent papers. you can't fake research fit
Most agents either run fully autonomously or interrupt at the wrong times.
What if agents know when YOU want to step in?
🚀Introducing PlowPilot - a web agent that adapts to your interaction patterns
achieving +26.5% user-reported usefulness
Huge credit to @FariaHuqOaishi for leading this project!
🚀 I'm hiring 3 postdocs to work with me @Krafton_AI (Seoul, Korea)
What I offer:
🔥 Access to a compute cluster with 1,000+ GPUs
🧠 Support to pursue top-tier research
💰 The highest postdoc salary in Korea
+ The absolutely best company-provided food & office view 😁
@SharonYixuanLi Awesome! Using AI is inevitable, even for conference reviews. I expect this decision to change a lot of things. And I’m sure new rules will be established along the way, like 'White prompts'