Check out the new CFP for GenBench. Benchmarking generalization for NLP tasks is more important than ever!

Come and join us within Naila Murray’s org who leads the foundational and exploratory research in Europe!

Apply at:
lnkd.in/eC3RaWND
lnkd.in/e5R5-QJr

AI at Meta AI at Meta Meta Open Source

Happy to serve this groundbreaking community!

Really enjoyed this conversation with Sam Charrington, where we covered recent progress in computer vision research, as well as what to look for in 2024 and beyond ;)

So proud of this work from the Code Llama team! Check out the different code model flavours and sizes, permissively licensed and SoTA among open models ➡️ tinyurl.com/codellama

Happy to be releasing Code Llama! We've built it on Llama 2 and improved it for code use cases. In particular it supports infilling out of the box, and was trained with sequences up to 16k tokens.
Looking forward to what the community will build with it! 1/7

Happy to announce that, after a successful edition with ICLR 2023, ICLR has agreed to continue to participate in the joint Journal-to-Conference track for its future conferences! This follows a similar engagement by NeurIPS. A really nice milestone for this initiative!

Very exciting work on thoughtfully augmenting language model capabilities from Timo Schick et al. at Meta AI

If you've published a JMLR paper since the beginning of 2022 and would like to present it at ICLR 2023, remember that the deadline to submit a request to present at the Journal-to-Conference track is January 13th 2023!

More info here: iclr.cc/public/Journal…

Final call on **in-person** #WiML workshop at NeurIPS Conference #NeurIPS2022

**Please RT to help us!** :-)

- Reception is in Nov 27th. Registration link: eventbrite.com/e/women-in-mac… All genders welcome!

- In-person workshop is in Nov 28th. Registration link: …shop-2022-registration.eventbrite.com

Really enjoyed working on this with my awesome collaborators Roberta Raileanu, Danielle Rothermel, Vegard Mella, Tim Rocktäschel and heiner, and our fearless leader Eric Hambro! If offline RL, observation-based, or imitation learning is your thing, check it out!

Dungeons and Data: A Large-Scale NetHack Dataset

Paper: arxiv.org/abs/2211.00539
Link: nips.cc/virtual/2022/p…

- Eric Hambro Roberta Raileanu Danielle Rothermel Vegard Mella Tim Rocktäschel heiner Naila Murray

FAIR Paris is hiring a research engineer interested in working on the intersection of NLP and reinforcement learning. If this sounds exciting to you, please apply via the link below or feel free to reach out! metacareers.com/v2/jobs/108307…

Delighted to announce this opportunity to present your recent JMLR papers (published on or after 01/01/2022) to one of the 3 high-visibility conferences below. More 👇🏽

Don't forget to register to hear Naila Murray tell us all about how to train deep vision models when data is limited! Online seminar talk open to everyone.

Super excited to present the technical write up of the #NetHackChallenge21 at NeurIPS Conference 21.

arxiv.org/abs/2203.11889

Read about the winning solutions, and breakdowns of bot performance for symbolic bots and deep RL agents.

This year, symbolic bots crushed it!

1/n

So happy to see this come out. I had a blast co-editing this book with Bogdan Ionescu and Wilma Bainbridge, and reading about fascinating recent advances in our understanding of how humans perceive visual stimuli. Worth checking out!

Thanks to Zoë Papakipos & joanna 👩🏻‍💻, the team of Meta AI engineers behind AugLy (github.com/facebookresear…), for the collaboration.

4/4

📊 Robustness Reports on ImageNet!

We've indexed AugLy's robustness reports, which show model vulnerabilities to different manipulations. Check it out below:

paperswithcode.com/sota/image-cla…

1/4

#CVPR2022 -- New Deadline.
11:59am (noon) PDT 11/18/2021