Reposting our 2025 ICML paper on the algorithmic evaluation of generative AI! #ICML2025

Reviewing of applications will begin on Monday, Jan 19. #FAccT #FAccT2025 #CHI2025 #ACL2025 #ICML2025 #NeurIPS2025 #FAccT2026 #CHI2026 #ACL2026 #ICML2026 #NeurIPS2026

Policy design for two-sided platforms with participation dynamics: Interview with Haruka Kiyohara - ΑΙhub

In this blog post, Haruka Kiyohara writes about work presented at #ICML2025, investigating the participation dynamics in two-sided markets. Find out more about the long-term impacts of decision-making algorithms on such two-sided platforms.
aihub.org/2025/10/09/p...

#ICML2025 outstanding position paper: Interview with Jaeho Kim on addressing the problems with conference reviewing - Robohub

#ICML2025 outstanding position paper: Interview with Jaeho Kim on addressing the problems with conference reviewing

by @aihub.org

robohub.org/icml2025-out...

Personalized Diffusion Models May Leak Fine‑Tuned Training Data

FineXtract can extract roughly 20% of images from fine‑tuned diffusion models; the study, presented at ICML 2025, notes that checkpoints often appear on Civitai and HuggingFace. getnews.me/personalized-diffusion-m... #finextract #icml2025

Understanding CLIP’s Domain and Compositional Generalization

Broad training-domain diversity is crucial for CLIP’s zero‑shot domain and compositional generalization, the ICML 2025 Spotlight study finds. Submissions ran Feb 2025–Sep 2025. Read more: getnews.me/understanding-clips-doma... #clip #zeroshot #icml2025

#ICML2025 outstanding position paper: Interview with Jaeho Kim on addressing the problems with conference reviewing At this year’s International Conference on Machine Learning (ICML2025), Jaeho Kim, Yunseok Lee and Seulki Lee won an outstanding position paper award for their work Position: The AI Conference Peer Review Crisis Demands Author Feedback and Reviewer Rewards. We hear from Jaeho about the problems they were trying to address, and their proposed author feedback mechanism and reviewer reward system.

#ICML2025 outstanding position paper: Interview with Jaeho Kim on addressing the problems with conference reviewing

At this year’s International Conference on Machine Learning (ICML2025), Jaeho Kim, Yunseok Lee and Seulki Lee won an outstanding position paper award for their work Position: The AI…

Memory traces in reinforcement learning - ΑΙhub

In our latest blog post, @onnoeberhard.com writes about work presented at #ICML2025 on partially observable reinforcement learning which introduces an alternative memory framework - “memory traces”.
aihub.org/2025/09/12/m...

#ICML2025 outstanding position paper: Interview with Jaeho Kim on addressing the problems with conference reviewing - ΑΙhub

How do you solve a problem like conference reviewing? Jaeho Kim explains all in this interview covering the position paper that won Jaeho and colleagues an outstanding paper award at #ICML2025 @icmlconf.bsky.social
aihub.org/2025/09/02/i...

6/ This work appeared earlier as a poster at ICML 2024 (ML4LMS workshop). Thanks to that community for early feedback.
openreview.net/forum?id=gGn...
#ICML #icml2025

The value of prediction in identifying the worst-off: Interview with Unai Fischer Abaigar - ΑΙhub

We hear from Unai Fischer (@lmumuenchen.bsky.social), one of the winners of an #ICML2025 outstanding paper award, about his work (with Christoph Kern and Juan Carlos Perdomo) on “The value of prediction in identifying the worst-off”

aihub.org/2025/08/27/t...

Allen School researchers develop machine learning technique to capture the chatter between brain regions - Allen School News Understanding how different parts of the brain communicate is like trying to follow conversations at a crowded party. Neuroscientists face a similar challenge: even when they can record signals from m...

Understanding how regions of the brain communicate is like trying to follow conversations at a crowded party. A team led by #UW #UWAllen prof @mattgolub.bsky.social recently made headway with a new machine learning method that cuts through the noise. #ICML2025 news.cs.washington.edu/2025/08/12/a...

Dare I say my personal computer is finally complete now? 😅 #icml2025

That’s my little #ICML2025 convex RL roundup!

If you know of other cool work in this space (or are working on one), feel free to reply and share.

Hope to see even more work on convex RL variations 🚀

n/n

We’re training AI on everything that we know, but what about things that we don’t know?

At #ICML2025, the "Exploration in AI Today (EXAIT)" Workshop sparked a crucial conversation: as AI systems grow more powerful, they're relying less on genuine exploration and more on curated human data.

it's extra work, but i really enjoy making creative posters.
these are the two i made for #ICML2025 last week.

i just sent another poster to get printed with a design i'm excited to hang on the poster boards @rl-conference.bsky.social next month! 😃

BC nature 💚💚💚 #icml2025

List of work presented at ICML 2025 by the Calgary ML Lab

@calgaryml.bsky.social was at #ICML2025 in Vancouver🇨🇦 last week: our lab had 6 different works being presented by 5 students across both the main conference & 6 workshops!

We are also recruiting fully-funded ML PhD/MSc students🧑‍🎓If interested, reach out to me with a resume ✉️

Our paper from the ICML 2025 workshop on Multi-Agent Systems explores a cascaded LLM framework for cost-effective decision-making. Small models defer to larger ones—or abstain for human input—based on uncertainty.

Paper: arxiv.org/abs/2506.11887
#LLM #MultiAgent #ICML2025

Our paper from the ICML 2025 workshop on Models of Human Feedback (Oral presentation) proposes few-shot steerable alignment for LLMs, utilising neural processes to handle conflicting and unobserved preferences.

Paper: arxiv.org/abs/2412.13998

#LLM #AIAlignment #ICML2025

Zebra: In-Context Generative Pretraining for Solving Parametric PDEs Solving time-dependent parametric partial differential equations (PDEs) is challenging for data-driven methods, as these models must adapt to variations in parameters such as coefficients, forcing ter...

Leveraging LLMs' in-context learning for modeling physical dynamics. Paper #ICML2025.
"Zebra: In-Context Generative Pretraining for Solving Parametric PDEs", Louis Serrano, Armand Kassai, Thomas Wang, Pierre ERBACHER, Patrick Gallinari
📄 arxiv.org/abs/2410.03437
🖥️ github.com/LouisSerrano...

CLAN for AI Research on Language and Networks CLAN for AI Research on Language and Networks, or **CLAN** for short, is a research group and a part of the larger [DIS Group](https://cs.au.dk/research/data-intensive-systems) in the [Department of C...

Back from #ICML2025 🛬, and off 🚄 to Norrköping 🇸🇪 for #ic2s2

CLAN (cs.au.dk/~clan/) members are presenting 2 papers: 1 spotlight and 1 oral. See 🧵for posters and summaries

👋 Reach out to chat about observational studies, causality, LLM agents, human-centered AI etc.

Enjoying the last minutes of #ICML2025

Wish LM could plan—not just guess the next word?
TRACE lets LM see all endings before each move.
– Global control at inference time
– Tractable lookahead via an HMM LM-proxy
– Linear classifier per constraint
Outperform RL, DPO, FUDGE—at just +20% decoding over base LM.
#ICML2025 @guyvdb.bsky.social

Final #ICML2025 lunch with Bernadette Bucher’s (UMich) group.

That’s a wrap! Had an AMAZING time, met so many great people (including many I’d only interacted with on social media), and (for once) actually explored an awesome city too. Next stop: home, Toronto.

Computational complexity. There are many problems that cannot be solved in constant or linear time. No amount of data can overcome this. H/t Dale Schuurmans at #ICML2025

Great to present what’s coming next for NDIF at the @actinterp.bsky.social workshop at #ICML2025!

If you missed us, let’s chat after the conference. Reach out here: forms.gle/LtTyYnkaxDyg...

Image from article in Radiology: Artificial Intelligence

CoRaX #AI system helps radiologists detect abnormalities on chest radiographs https://doi.org/10.1148/ryai.240277 @uh.edu #ICML2025 #MachineLearning #AI

Found my community here. LLMs and MLLMs for Life Sciences. @icmlconf.bsky.social #icml2025

Sergey Levine was just presenting in the Exploration in AI @ #ICML2025 and promoted that exploration needs to be grounded, and that VLMs are a good source ;-) Check our paper below
👇