KAIST @ ICML satellite event happening right now with great panel discussions 🔥

Excited to share RoboWorld 🤖 We roll out generalist robot policies from 4,186 real initial scenes, entirely inside a video world model with no robots, and the rankings hit Pearson r = 0.989 (Spearman ρ = 0.970) with the real RoboArena leaderboard. 🧵 [1/7]

🌟 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) luma.com/c3e67mf0 Seong Joon Oh (@coallaoh) — Organizer ProbML 2026 (7/5) probml.cc/ Sungsoo Ahn (@sungsoo_ahn_) —Organizer Juho Lee — Keynote Speaker To be held at KAIST Hongneung Campus KAIST @ ICML 2026 (7/6) kaist-icml.github.io/ - 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) holymollyhao.github.io/ai4sc… Sungsoo Ahn (@sungsoo_ahn_) — Organizer (EXPO Talk Panel) Seoul World Model: Grounding World Simulation Models in a Real-World Metropolis (7/6) icml.cc/virtual/2026/75730 Seungryong Kim (@KimSeungry62571) — Organizer WiML Symposium @ ICML 2026 (7/8) wiml.org/events/wiml-symposi… Hyunjung Shim — Organizer ICML SPIGM Workshop (7/10) spigmworkshop2026.github.io/ Seul Lee (@SeulLee05) — Organizer ICML AI4Science Workshop (7/11) ai4sciencecommunity.github.i… Sungsoo Ahn (@sungsoo_ahn_) — Organizer ICML Philosophy Meets Machine Learning Workshop  (7/11) sites.google.com/view/philml… Jaesik Choi — Organizer ICML Statistical Frameworks for Uncertainty in Agentic Systems Workshop (7/11) agentic-uncertainty-icml2026… 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!

SUPER excited to share this work!! From what we know, this the first case of AI Co-Scientist developing a competitive generative deep-learning algorithm for a mature scientific ML benchmark. Big congrats to @rldudrldbs2 for taking the risk on this unexplored type of research!

We built a Human–AI co-discovery system (HACO) that discovered MaskGXT, a new SOTA algorithm for crystal structure prediction (CSP)! 💎🤖 While agentic research systems like Karpathy's AutoResearch focus on refining a fixed method, we aimed at finding a new generative principle.

They plan to begin accepting motivated students starting from the Fall 2026 semester. If you're interested in joining their labs, please feel free to contact them! Sehoon Kim: sehoonkim.org/ Hyunwoo Kim: hyunw.kim/ Seung Wook Kim: seung-kim.github.io/seungkim…

📢 Three incoming faculty members at KAIST AI, starting in August 2026✨ Dr. Sehoon Kim from xAI (@sehoonkim418), Dr. Hyunwoo Kim (@hyunw_kim), and Dr. Seung Wook Kim (@seungkim0123), both from NVIDIA, will be joining KAIST AI as Assistant Professors Check their websites below🧵

Can a robot understand the nonverbal signals you give in real time — your pointing gestures, your gaze, the things you never put into words? Meet EDITH: a framework that lets robots comprehend and act on human nonverbal signals. project-edith.github.io 🧵[1/n] @KAIST_AI #Robotics #HumanRobotInteraction #VLA #ProjectAria

The reversal curse. Edits that don't suppress negations. Multi-hop updates that don't propagate. These look like separate bugs. Our ICML 2026 spotlight argues they may share a common geometric origin, visible only when you study how representations move under updates 🧵 (1/11)

📣Join us at the Global AI Frontiers Symposium 2026 in Seoul, right before ICML🌏🌟 Featuring keynotes from Leslie Pack Kaelbling (@MIT_LISLab) and Noam Brown (@polynoamial), plus panel discussions with Kyunghyun Cho (@kchonyc) and Emily Black! aifrontiers.kr/

Happy to share that our work, Reward Score Matching (RSM), has been selected as an Oral Presentation at the SPIGM Workshop at ICML 2026. RSM asks a simple question: The literature on RL fine-tuning for diffusion/flow models looks fragmented, but which differences are actually fundamental? 🔗 arxiv.org/abs/2604.17415

🚨Most AI agents solve only the problems users explicitly ask about. But what about the problems users haven’t noticed yet? 🌊TIDE enables proactive multi-problem discovery, helping agents uncover hidden issues 🔍 and recommend actionable next steps ✅. huggingface.co/papers/2606.0…

Happy to share our #Interspeech2026 paper!🗣️ arxiv.org/abs/2509.17901 w/ @seo_minjoon @KAIST_AI #NAVERCloud Quite a few video-LLMs still process video muted. Auditing 10 benchmarks, we find heavy visual shortcuts. We then make listening practical by compressing audio tokens 25×

Can MLLMs actually track what's happening in a video? Introducing VSTAT 🎯, our new benchmark for visual state tracking. The tasks are simple: count cups, read typed words, count page flips. Humans solve them easily. MLLMs don't. vision-x-nyu.github.io/vstat… 🧵 [1/11]

A new RLHF vulnerability identified 🚨 RLHF can be exploited to optimize misaligned biases, such as ideological or promotional biases. We introduce Alignment Tampering, a vulnerability where the LLM undergoing alignment influences the preference dataset itself, causing RLHF to amplify undesired behaviors. 💻 Paper & Code: alignment-tampering.github.i… #ICML2026 #AIAlignment @KAIST_AI, @MIT_CSAIL 1/N 🧵

What if your retriever could speak every language your data speaks? 🌐 Your answer might live in a document 📄, a SQL table 🗃️, an RDF knowledge graph 🔗, or a property graph 🕸️, and OmniRetrieval reaches into all of them, meeting each source in its own native query language instead of flattening everything into one lossy space. Paper: huggingface.co/papers/2605.2…

Excited to introduce 🧑‍🎓𝗟𝗲𝗮𝗿𝗻 𝗳𝗿𝗼𝗺 𝗪𝗲𝗮𝗸𝗻𝗲𝘀𝘀𝗲𝘀 (LearnWeak)! A framework that automatically specializes small CUAs for specific domains by 🎯𝘁𝗮𝗿𝗴𝗲𝘁𝗶𝗻𝗴 𝘁𝗵𝗲𝗶𝗿 𝗼𝘄𝗻 𝗳𝗮𝗶𝗹𝘂𝗿𝗲 𝗽𝗮𝘁𝘁𝗲𝗿𝗻𝘀 in data generation and training. 🧵(1/7)

🚀 Releasing ✨AXPO✨ an RL method to lift agentic reasoning models past their next scaling tier. Be it math, perception, or search, AXPO fixes the structural blind spot 'just add tools' recipes leave untouched. 8B beats 4x larger 32B baseline on Pass@4. from NVIDIA 🧵 (1/7)

Introducing TRQAM! Internalizing a KL trust region inside the sampling SDE stabilizes off-policy RL fine-tuning of pretrained flow policies. With TRQAM, we lift offline RL success on 50 OGBench tasks from 46% to 68%. 🧵 [1/8] yonghdong.github.io/blog/trq…

🚨New Optimizer Paper AMUSE: Anytime MUon with Stable gradient Evaluation AMUSE combines Muon with Schedule-Free-style gradient evaluation for stable anytime training without LR decay. • Stronger 124M / 720M / 1B pretraining • Strong ImageNet / ViT fine-tuning performance.