Oscar Mañas's Avatar

Oscar Mañas

@oscmansan.bsky.social

Visiting researcher at Meta FAIR, PhD candidate at Mila and Université de Montréal. Working on multimodal vision+language learning. Català a Montreal.

164 Followers  |  172 Following  |  9 Posts  |  Joined: 24.11.2024  |  1.404

Latest posts by oscmansan.bsky.social on Bluesky

Headed to @cvprconference.bsky.social in Nashville! I'll be presenting our work on Multimodal Reward-guided Decoding. Let's connect if you're around!

10.06.2025 18:20 — 👍 2    🔁 0    💬 0    📌 0

TFW you find a memory leak in your code two days before the rebuttal's deadline

15.05.2025 14:17 — 👍 1    🔁 0    💬 0    📌 0

Heading to Singapore for the next 1.5 weeks for @iclr-conf.bsky.social. If you're around and want to meet up, hit me up!

21.04.2025 23:49 — 👍 1    🔁 0    💬 0    📌 0
Post image Post image Post image Post image

"Tokenize Everything!" Luke Zettlemoyer of
@uofwa.bsky.social on using GPT-like autoregressive techniques for training multimodal models (text, images, audio etc.) at the Simons Institute workshop on The Future of Language Models and Transformers simons.berkeley.edu/workshops/fu...

01.04.2025 20:55 — 👍 4    🔁 1    💬 0    📌 0

Yes! I always recommend this book to fellow researchers for good coding practices :)

30.12.2024 09:26 — 👍 2    🔁 0    💬 0    📌 0
Gemini 2.0 and the evolution of agentic AI with Oriol Vinyals
YouTube video by Google DeepMind Gemini 2.0 and the evolution of agentic AI with Oriol Vinyals

I quite like this analogy by Oriol Vinyals:
* LLM ~= core electric brain
* Agent ~= LLM with a digital body

youtu.be/78mEYaztGaw

22.12.2024 01:32 — 👍 3    🔁 0    💬 0    📌 0
Post image

Curious about how to effectively steer the behavior of multimodal LLMs during inference to improve their visual grounding?

Join me today at 4:30pm at the AFM workshop at @NeurIPSConf, where I'll be presenting a poster on my work. Come by to learn more!

openreview.net/forum?id=VWJ...

14.12.2024 18:18 — 👍 3    🔁 1    💬 0    📌 0
Preview
Controlling Multimodal LLMs via Reward-guided Decoding As Multimodal Large Language Models (MLLMs) gain widespread applicability, it is becoming increasingly desirable to adapt them for diverse user needs. In this paper, we study the adaptation of...

Tomorrow at 3:15pm I'll be presenting my work at @mila-quebec.bsky.social's booth (#104) at @neuripsconf.bsky.social. Come to learn more about controlling multimodal LLMs via reward-guided decoding!

🔗 openreview.net/forum?id=VWJ...

10.12.2024 03:04 — 👍 11    🔁 3    💬 0    📌 0

All this being said, Meta/FAIR remains the only place where you can do open AI research with a group of stellar colleagues ten times larger than any university + big-tech computational capabilities level.

05.12.2024 17:48 — 👍 40    🔁 1    💬 2    📌 0

Oh I love the idea! I'll be happy to contribute :)

30.11.2024 23:04 — 👍 1    🔁 0    💬 0    📌 0

A bit late to the party but here we are!

27.11.2024 07:24 — 👍 2    🔁 0    💬 0    📌 0

@oscmansan is following 20 prominent accounts