🔥 A sneak peek at the incredible #ICOMP2026 lineup! Here is a subset of our confirmed plenary speakers so far:
🎙️ Michael Jordan 🎙️ Anatoli Juditsky 🎙️ @SebastienBubeck (online) 🎙️ @Ana_koloskova 🎙️ @konstmish 🎙️ @orvieto_antonio
More stellar speakers to be announced soon!
Replying to @ed_gorbunov
We're excited to welcome the following confirmed plenary speakers: Sebastian Bubeck (@SebastienBubeck, online), Anatoli Juditsky, Michael Jordan, Anastasia Koloskova (@Ana_koloskova), Konstantin Mishchenko (@konstmish), and Antonio Orvieto (@orvieto_antonio).
Martin Takac retweeted
📣 Call for Papers — ICOMP 2026
3rd International Conference on Computational Optimization
📍 Casablanca, Morocco · 3-6 November 2026
New Deadlines:
• Abstract: 20 June 2026
• Full paper: 5 July 2026
🔗 Submit: openreview.net/group?id=ICOM…
🌐 icomp.cc
Martin Takac retweeted
Scaling RL to long horizons remains a major challenge.
Long-horizon Q-learning (LQL) prevents compounding bootstrapping errors by bounding the difference in value over long horizons.
It shows large gains over 1-step TD and n-step returns!
Paper: arxiv.org/abs/2605.05812
In RL, what if we could learn from any experience from any policy in a way that is reliable and scalable?
This would be helpful in domains like robotics where new data is expensive.
We introduce Long-horizon Q-learning (LQL) to tackle this arxiv.org/abs/2605.05812.
Martin Takac retweeted
🚨 New Paper 🚨
ScheduleFree+: Scaling Learning-Rate-Free & Schedule-Free Learning to Large Language Models
A few modifications to Schedule-Free Learning make it completely LR tuning free, and allow it to greatly outperform schedules for long duration training!
arxiv.org/abs/2605.19095v1
MBZUAI is proud to be part of Spring School - AI For Impact, convening the brightest minds in AI to tackle real-world challenges in social impact.
Day 1 – March 23
🎙️ Round Table: AI Foundations (11:15 – 12:15) Moderated by MBZUAI's Department Chair and Professor of Machine Learning Eric Moulines, featuring Assistant Professor of Machine Learning Salem Lahlou and Associate Department Chair and Associate Professor of Machine Learning Martin Takáč alongside leading voices from University of California, Berkeley, Google DeepMind, Inria, and CentraleSupélec.
Day 2 – March 24
🔬 GFlowNets for Diverse Generation (09:45 – 10:30) Salem Lahlou
⚙️ From Distributed Optimization to Federated Learning (10:30 – 11:15) Martin Takáč
Day 3 – March 25
🧠 AI Foundation Models (14:00 – 15:00)- MBZUAI President and University Professor Eric Xing
⚡ Debate: World Models (15:00 – 16:30) - Eric Xing & Yann LeCun
More information here: ai4impact2026.com/
#MBZUAI #AIForImpact #ArtificialIntelligence #AIResearch
Martin Takac retweeted
Impressive position of the 5-year old @mbzuai of UAE in this picture 😊 Congratulations to our faculty and students!
NeurIPS 2025's top 50 paper contributors show China and the US neck-and-neck.
In the US, corporate labs—Google DeepMind, Meta, Microsoft—now rival top schools like Stanford, CMU, MIT, showing a major shift to industry-led AI research.
China’s leaders remain academic: Tsinghua, CAS, Peking, SJTU, HKUST. Singapore (NUS, NTU), Korea (KAIST), UAE (MBZUAI), and Canada (Mila) hold strong.
Europe trails, with only Oxford, EPFL, ETH Zurich, and TUM from the EU-27 making the list.
---
Source:
linkedin. com/posts/pierre-alexandre-balland-20b75b13_who-pushed-the-ai-frontier-at-neurips-2025-activity-7403119036496162817-vRPE
Martin Takac retweeted
Call for participation:
KAUST Workshop on Distributed Training in the Era of Large Models
kaust.edu.sa/events/dtelm25/
location: KAUST, Saudi Arabia
dates: Nov 24-26, 2025.
There will be a chance for some participants to present a poster and/or give a lightning talk.
Martin Takac retweeted
Struggling with data heterogeneity in your deep learning projects? We introduce ALSO, a new, practical Distributionally Robust Optimizer (DRO)🚀
📜 Paper thread for "Aligning Distributionally Robust Optimization with Practical Deep Learning Needs"
Our paper introduces the first Quasi-Newton method with global convergence as fast as - and even faster than - gradient descent!
✅O(1/k) global rate
⚡Accelerated version: O(1/k²)
🧠Adaptive, cubic regularization for stability & speed
Quasi-Newton finally catches up with theory.
1
3
142
Martin Takac retweeted
🔬 The lab of the future is here.
A place where robots run the experiments.
AI makes sense of the data.
Scientists steer the breakthroughs.
A new Science Robotics article co-authored by #MBZUAI’s VP for Research Prof. Sami Haddadin shows how this could cut years of discovery into months.
mbzuai.ac.ae/news/ai-and-rob…
#AI #Robotics #LabAutomation @LivUni @UofT @TU_Muenchen
🎉 Accepted to #NeurIPS2025 (Benchmark Track)!
SVRPBench: a realistic benchmark for stochastic VRP—traffic peaks, log-normal delays, accidents, and real time windows—across 500+ urban-scale instances.
Learn more: arxiv.org/abs/2505.21887
#MBZUAI #VRP #OR
10
260
🚀 AI4OPT Challenge!
Build an agentic AI that reads, experiments, and writes a mini-paper.
Tweak the pipeline. Beat the baselines. Win an award.
Check: ai4opt.fmin.xyz
#AgenticAI #ML #Optimization #LLM #MBZUAI #ICOMP
Martin Takac retweeted
The upcoming launch of the "K2 Think" model will mark a significant step in advancing artificial intelligence from the
UAE to the world, reflecting our leadership's vision for technological progress. K2 Think will combine the efficiency of smaller models with world-class performance and superior inference speeds, surpassing larger systems to set a new global benchmark for open-source reasoning
Developed through the partnership of @MBZUAl and @G42ai, it showcases the strength of cooperation between academia, public institutions, and the private sector in turning national aspirations into reality.
We congratulate the UAE and its leadership on this achievement, as our nation continues to enhance its global standing across all fields.
New Quasi-Newton method CEQN offers a simple, explicit stepsize with guaranteed global convergence for convex optimization, matching accelerated methods’ rates even with inexact Hessians. #Optimization #MBZUAI arxiv.org/pdf/2508.19712
1
173
Join the workshop at #icml25 – at Vancouver, British Columbia
We are #hiring for MBZUAI Postdoctoral Opening—Graph AI for Materials & Catalysis at MBZUAI. Message me if you're interested in joining our team. We are attending International Conference on Machine Learning 2025 if you would like to meet! - via #Whova event app – at Vancouver, British Columbia
3
285
🔧 Comp-opt friends: show us your solver magic!
#ICOMP2025 lands in Abu Dhabi 🇦🇪.
Abstracts due Aug 1, papers Aug 15 (AoE). Expect killer talks + sunny networking.
Details/CFP ⬇️
lnkd.in/gEapDXwB
#CallForPapers #Optimization #ML
3
137























