CS PhD student @UofT and @VectorInst. Ex research intern @Apple ML Research. Interested in deep learning theory and generative modeling.
Toronto, Ontario
Joined October 2024
- Tweets74
- Following277
- Followers457
- Likes619
Pinned Tweet
Can RL with outcome rewards alone just efficiently explore outside the support of the base model and learn completely new capabilities?
We study it in a theoretically tractable setup, and prove that outcome rewards are not enough, but process rewards can get you there! [1/3]
Alireza Mousavi retweeted
Standard PG analysis requires "all-iterate coverage" ie bounded π*/π_t ∀𝐭. What if we only assume it at 𝐭=𝟎 (eg base model has nontrivial success rate on all prompts)?
Ans: 𝐞𝐱𝐩𝐨𝐧𝐞𝐧𝐭𝐢𝐚𝐥 complexity due to "coverage collapse". arxiv.org/abs/2611.14467
Lol maybe this is what superintelligence is after all. We just can't understand it (and probably shouldn't, by definition) and disregard it as slop?
What should I do if some ICLR submissions are so much of a slop that I can't even write a summary about them? Sometimes there is not a single sentence in the submission that I think would make sense to a human. Quite funny artifact of the RLVR pressure on LLMs.
1
4
1,111
What should I do if some ICLR submissions are so much of a slop that I can't even write a summary about them? Sometimes there is not a single sentence in the submission that I think would make sense to a human. Quite funny artifact of the RLVR pressure on LLMs.
Alireza Mousavi retweeted
We have now crossed the threshold where developing theory has become orders of magnitude faster. It is now possible to rapidly develop theory that can drive AI research, not just explain it post-hoc.
For those who like deep learning "theory", our job for now seems to be finding the setting with the right amount of generality/complexity where models can produce meaningful (to humans) mathematical insights. How long till the model itself develops the taste to find it?
7
313
I’m at ICML 🇰🇷 presenting a 🔦 spotlight 🔦 poster today. Check it out if you want to know how process rewards help RLVR go beyond the base model support when outcome rewards alone can’t.
📆 Thu. 5-6:45 pm. Hall A. Poster # 1307.
Alireza Mousavi retweeted
🚀 🚀 🚀 Excited to share our new paper:
Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration
What does it take for an agent to stay curious in a 3D world?
The answer is memory.
🌐 Project: recuriosity.github.io/
📄 Paper: arxiv.org/abs/2605.22814
💻 Code: github.com/recuriosity/recur…
Under the spotlight at ICML 🔦 Take a look at what theory tells us about RLVR and base models.
Can RL with outcome rewards alone just efficiently explore outside the support of the base model and learn completely new capabilities?
We study it in a theoretically tractable setup, and prove that outcome rewards are not enough, but process rewards can get you there! [1/3]
28
4,223
How do you choose which mixer to use at each layer? Why not use all of them? Check out this work.
(1/8)
🚀 Introducing Super Apriel: One Checkpoint, Many Speeds
Train once → serve at any speed-quality tradeoff
We release:
✓ 15B supernets with 4 mixers/layer
✓ Training code (Fast-LLM)
✓ vLLM serving extension
🧵 How it works ↓
3
296
Alireza Mousavi retweeted
Attention and MLP matrices can store as much knowledge as # params, but in practice, finite samples/compute and long context noise can further limit this capacity. We characterize these trade-offs in a needle-in-a-haystack model.
#ICLR2026 poster Thurs, led by @nurimertvural45
Alireza Mousavi retweeted
Excited to share @YaldaForoutan and @pekzta4's work on robust 3D reconstruction from 360 captures!
Casual 360 video capture gives you the coverage you need for calibration and reconstruction; FullCircle makes it hassle-free!
No more wrestling with stitching artifacts or floaters from the camera operator. Data and code are released, enjoy!
📢📢📢 Introducing "FullCircle: Effortless 3D Reconstruction from Casual 360° Captures"
TL;DR: 10x faster casual capture with clean reconstructions
Homepage: theialab.github.io/fullcircl…
Code: github.com/theialab/fullcirc…
arXiv: arxiv.org/abs/2603.22572
Led by Yalda Foroutan & Ipek Oztas
We prove the number of reward queries in RLVR depends on the “likelihood quantile” function of the base model.
Unfortunately, LQ becomes exponentially small if we want to go significantly below the base model error, then RL needs exponentially many reward queries. [2/3]
This is no longer an issue if you use process rewards. Number of reward queries will depend on a “token-level likelihood quantile” that is never exponentially small in sequence length. [3/3]
📄 arXiv: arxiv.org/abs/2603.06957
Alireza Mousavi retweeted
New paper 🥳 RL relies a lot on an agent’s capability to explore. Our strategy-guided exploration makes the agent find new solutions more efficiently. It learns faster, and in some environments its Pass@1 surpasses the base model’s Pass@128. 🧵1/6
📄 arxiv.org/abs/2603.02045
Alireza Mousavi retweeted
Masked Diffusion LMs (MDLMs) are the most exciting paradigm shift in AR generation because they can decode in parallel, infill, and self-correct.
But they are bottlenecked by the transformer's quadratic attention, making throughput fall apart for long contexts.
We offer a simple solution. Introducing DiffuMamba: first diffusion LM with a bidirectional Mamba backbone. Better quality. Up to 8.2x faster.
🧵1/N
Alireza Mousavi retweeted
Fantastic workshop on High-Dimensional Learning Dynamics at BIRS. All talks here:
birs.ca/events/2026/5-day-wo…
Now accepted at ICLR 2026. Check it out if you’re a diffusion/flow/OT enjoyer.
Super excited to share what @stephenz_y and I’ve been up to during our internship at🍎:
Using optimal transport makes flows straighter and generation faster in flow matching, but small batch OT is biased and large batch OT is slow.
What to do? Use semidiscrete OT!
🧵
Alireza Mousavi retweeted
- Sat Dec 6 @ SPIGM workshop, On Fitting Flow Models with Large Sinkhorn Couplings, work with @alirezamh_ , Michal Klein, Marco Cuturi 4/4
Alireza Mousavi retweeted
Our research team is hiring PhD interns 🍏 Spend your next summer in Paris and explore the next frontiers of LLMs for uncertainty quantification, calibration, RL and post-training, and Bayesian experimental design.
Details & Application ➡️ jobs.apple.com/en-my/details…