
Sarath Chandar's research group at @polymtl, @UMontreal and @Mila_Quebec focusing on Machine Learning!
Montrรฉal, Quรฉbec, Canada
Joined June 2021
- Tweets182
- Following147
- Followers749
- Likes110
Chandar Lab retweeted
Just got accepted at NeurIPS! Read our paper on how (we think) TFMs generalize!
arxiv.org/pdf/2608.17957
Chandar Lab retweeted
Our semantic world models paper is accepted at #NeurIPS2026 E&D Track! ๐
What's the best latent space for diffusion world models? We find semantic spaces beat VAE-style ones on planning and robustness, with competitive visual quality.
Paper & more: hskalin.github.io/semantic-wโฆ
Diffusion world models can help test and improve robot policies before running them on real robots.
But can the choice of latent space make the WM more faithful?
We show that semantic spaces beat reconstruction spaces on task relevant metrics.
hskalin.github.io/semantic-wโฆ
Chandar Lab retweeted
Read our paper on scaling interpretable LLMs, we show that interpretable LLMs are not only possible but can scale both on typical generation benchmarks and interpretability benchmarks. arxiv.org/abs/2608.07594๐
Chandar Lab retweeted
People use LLMs all the time for help with medical problems, some doctors might also use them too to explain diagnoses to patients.
But diagnostic tests are inherently probabilistic. This work from @dicemasz explores if LLMs are suited or not for this usecase!
Chandar Lab retweeted
The goal of this program is to identify under represented students from very early stage (high school) and introduce them to CS and programming! Iโm super proud of this initiative that is getting mature over the last 3 years!
Chandar Lab retweeted
We are at ICML 2026! ๐ฐ๐ท
๐
Date: Friday, July 10, 2026
๐ Room 403, COEX, Seoul
๐๏ธ Workshop Schedule: icml.cc/virtual/2026/workshoโฆ
๐ More Details: weightsymmetry.com/
Chandar Lab retweeted
At #ICML2026 this week! Happy to chat about LLM training dynamics, model adaptation, and model merging.
We'll also be hosting the Workshop on Weight-Space Symmetries on Friday โ come say hi! ๐
@Mila_Quebec
We are at ICML 2026! ๐ฐ๐ท
๐
Date: Friday, July 10, 2026
๐ Room 403, COEX, Seoul
๐๏ธ Workshop Schedule: icml.cc/virtual/2026/workshoโฆ
๐ More Details: weightsymmetry.com/
Chandar Lab retweeted
@ChandarLab will be at #ICML 2026
in Seoul โ South Korea ๐ฐ๐ท !! Presenting our most recent state-of-the-art research on continual learning, reinforcement learning and AI agents! Drop by and check it out! ๐ฅ
Chandar Lab retweeted
We are super excited to announce @ChandarLab 7th Annual Research Symposium! ๐ฅ
Join us on July 23โ24 for two days of talks on deep learning, NLP, reinforcement learning, continual learning, and AI for science, plus a keynote by @mengyer from NYU!
Event is hybrid!
Chandar Lab retweeted
Extremely excited to share our recent work on diffusion world models. We ask a simple question - what space supports diffusion world modeling the most and how do we evaluate that?Turns out representation is the answer with JEPA space yielding the strongest diffusion world models!
Diffusion world models can help test and improve robot policies before running them on real robots.
But can the choice of latent space make the WM more faithful?
We show that semantic spaces beat reconstruction spaces on task relevant metrics.
hskalin.github.io/semantic-wโฆ
Chandar Lab retweeted
Diffusion world models can help test and improve robot policies before running them on real robots.
But can the choice of latent space make the WM more faithful?
We show that semantic spaces beat reconstruction spaces on task relevant metrics.
hskalin.github.io/semantic-wโฆ
Chandar Lab retweeted
Still at #ICLR2026 and interested in training dynamics, simplicity bias, training under data distribution shift, and model merging? Come to our workshop posters to see what we are working on! ๐งต๐
Chandar Lab retweeted
Want to know the expressivity of Mamba 3?
Come by our ICLR poster!
Sat, Apr 25 โข 3:15 PM โ 5:45 PM
Pavilion 4 P4-#4409
The Expressive Limits of Diagonal SSMs for State-Tracking
Joint work with Behnoush Khavari, Siamak Ravanbakhsh, and @apsarathchandar.
Chandar Lab retweeted
I'm in Rio de Janeiro for ICLR 2026! ๐ง๐ท
I will be presenting our paper "Is Depth Heterogeneity a Barrier to Model Merging?" (spoiler: not really) at the Workshop for Test-time Updates on Monday April 27th!
Please dm if you're there and you'd like to geek out about... (1/2)
Chandar Lab retweeted
๐ฃ Announcing the CoLLAs Seminars
A year-long exploration of one of the central challenges in AI: building systems that can learn continually, adapt in real time, and improve over their lifetime.
Join us on May 13th at 11 am ET as we kick off the series with Pulkit Agrawal speaking on โRethinking Post trainingโ.
โน๏ธ Learn more: lnkd.in/erEdDxgP
โ๏ธ Join our mailing list: lnkd.in/eEGwH-3E
๐ Zoom link for the talk: lnkd.in/ekkHE5nX
Chandar Lab retweeted
If you work in continual learning, adaptation, online learning, updatable ML, unlearning, model editing and any other form of non-stationary/non-iid learning settings, consider joining the @CoLLAs_Conf mailing list here: groups.google.com/g/collas-lโฆ!
Chandar Lab retweeted
Hierarchical planning unlocks long-horizon, non-greedy behavior in JEPA world models.
Paper:ย arxiv.org/pdf/2604.03208
Website:ย kevinghst.github.io/HWM/
Code:ย github.com/kevinghst/HWM_PLDโฆ
Chandar Lab retweeted
The abstract deadline for @CoLLAs_Conf is this Friday! If you work in any non I.i.d learning settings, consider submitting to CoLLAs!
The CoLLAs submission deadline is this Friday!
We invite researchers to explore all facets of ML adaptation, from incorporating new capabilities during continuous training to efficiently removing outdated or harmful data.
- ๐๐ฏ๐๐๐ฟ๐ฎ๐ฐ๐ ๐๐ฒ๐ฎ๐ฑ๐น๐ถ๐ป๐ฒ: April 10, 2026
- ๐ฆ๐๐ฏ๐บ๐ถ๐๐๐ถ๐ผ๐ป ๐๐ฒ๐ฎ๐ฑ๐น๐ถ๐ป๐ฒ: April 15, 2026
- ๐๐ผ๐ป๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ ๐๐ฎ๐๐ฒ๐: Sep 14โ17, 2026
๐ ๐๐ผ๐ฟ ๐ณ๐๐น๐น ๐ฑ๐ฒ๐๐ฎ๐ถ๐น๐ ๐ผ๐ป ๐๐ต๐ฒ ๐๐ฎ๐น๐น ๐ณ๐ผ๐ฟ ๐ฃ๐ฎ๐ฝ๐ฒ๐ฟ๐: lifelong-ml.cc/Conferences/2โฆ
Chandar Lab retweeted
I wrote an article on agentic coding for beginners after my talk at @apsarathchandar @ChandarLab group. We cover history of AI coding tools, the importance of model harnesses, and general principles in simple research workflows.
Feedback is very welcome!
aidanli.dev/writing/articlesโฆ
Chandar Lab retweeted
I am attending EACL in Morocco! If you are going and want to chat, please let me know! I will present SafeSearch: a reinforcement learning approach for safe LLM agents, proposed by @AmazonScience's amazing intern @ZhanQiusi1 and our great team!
Paper๐: lnkd.in/e342i9WK

























