Abdelhakim Benechehab
@abenechehab
Followers
54
Following
54
Media
9
Statuses
35
Ph.D. Student @ Huawei Noah’s Ark Lab | Working on model-based reinforcement learning
Paris
Joined May 2023
🚀 I'm happy to share that our latest paper has been accepted at #ICML2025 🌟. 📌 "AdaPTS: Adapting Univariate Foundation Models to Probabilistic Multivariate Time Series Forecasting". 👉 paper & code: See you in Vancouver! 🍁.
1
4
10
RT @Cohere_Labs: Our ML Theory group is looking forward to welcoming @AmbroiseOdonnat next week on Thursday, June 19th for a session on "La….
0
5
0
Kudos to my amazing collaborators and supervisors: @Vasya2104 @_GPaolo @albertcthomas @MaurizioFilip19 @balazskegl ❤️. A detailed thread coming soon ⏳🧵.
0
0
2
🤝A joint work with: @youssef_attiaeh.@AmbroiseOdonnat.@oussamazekri_.@albertcthomas.@_GPaolo.@MaurizioFilip19.@IevgenRedko.@balazskegl .@Vasya2104 . I'll be around all week—feel free to reach out if you’d like to chat!.
0
0
3
TL;DR: We introduce adapters to enhance pre-trained univariate time-series foundation models (FMs) for multivariate tasks, addressing feature dependencies and uncertainty quantification.
1
0
1
📄 AdaPTS: Adapting Univariate Foundation Models to Probabilistic Multivariate Time Series Forecasting .📌 Monday April 28, 11:00 a.m. - 12:00 p.m. — Workshop on Adaptive Foundation Models (SCOPE@ICLR) .📍 Where: Peridot 204-205.🔗 Paper & Code:
1
1
1
TL;DR: We investigate in-context learning using LLMs for dynamics learning of Markov Decision Processes with applications in off-policy reinforcement learning.
1
0
1
📄 Zero-shot Model-based Reinforcement Learning using Large Language Models .📌 Saturday April 26, 09:00 a.m. - 11:30 a.m. — Main Conference Poster Session .📍 Where: Hall 3 + Hall 2B #375 .🔗 Paper & Code:
1
1
1
Excited to be heading to Singapore for #ICLR2025 this week! 🇸🇬 . I will be presenting our two latest works across the main conference and a workshop. 🧵.
1
4
10
Our latest work on multi-agent hierarchies led by @_GPaolo 🔥.
Living organisms organize as loosely coupled hierarchies—while our AI systems remain rigid monoliths. What if we built AI that scales and adapts like biological systems instead?.What would an AI like this look like? 🧵. 1/9.
0
1
6
RT @AmbroiseOdonnat: 🚀Proud to share our work on the training dynamics in Transformers with @_Vassim & @CabannesVivien @Inria @AIatMeta!….
0
8
0
🚀 I'm happy to share that our latest paper has been accepted at #ICLR2025 🌟. 📌 "Zero-shot Model-based Reinforcement Learning using Large Language Models". See you in Singapore! 🇸🇬.
Looking to leverage LLMs for multivariate time series forecasting? . 🎉 Search no more! You can do exactly that with our new package DICL (Disentangled In-Context Learning). 🖥️ Code & Demo: 📜 New preprint: DICL for RL . 1/🧵
1
4
14
RT @shaokun1991: [1/3] Happy to share our new GUI Agent paper: SPA-Bench: A Comprehensive Benchmark for Smartphone Agent Evaluation. Key co….
0
6
0
RT @martin_gorner: I'm looking at the "AI achieves Kaggle Grandmaster Level" paper published last week: A massive….
0
9
0
🔥🔥.
I am excited to present Agent K as the first end-to-end agent (i.e., autonomous from Kaggle URL to submissions that win competitions) to achieve an equivalent of Kaggle grandmaster level. Our agent codes the whole data science pipeline from a natural language description of the
0
0
1
RT @albertcthomas: Machine learning researchers—struggling with Python development behind a corporate proxy? In my latest blog post, I shar….
0
3
0
Check out the new episode of our podcast Aljadid Fl'AI, with the latest AI news brought to you in Moroccan Dialect 📢📢. Presented at @MathsMaroc , Sponsored by @adria_B_T .
0
0
1
RT @AdinaYakup: 🚀 Huawei AI Lab - Noah's Ark has been sharing more papers on the Daily Papers page @huggingface recently 📑 Kudos to @abene….
0
9
0
❤️Huge thanks to all my wonderful collaborators and supervisors at Huawei Noah's Ark Paris and Kaust: @youssef_attiaeh @AmbroiseOdonnat @oussamazekri_ @albertcthomas @_GPaolo @MaurizioFilip19 @IevgenRedko @balazskegl.
0
0
4
🍒 Cherry on top: We visualize attention layers of the LLaMA 3 8B model, showing that pre-trained LLMs can capture repeating time patterns through attention weights. 6/🧵
1
0
2