Huge congrats to AI at Meta on the Llama 3.1 release! Few notes: Today, with the 405B model release, is the first time that a frontier-capability LLM is available to everyone to work with and build on. The model appears to be GPT-4 / Claude 3.5 Sonnet grade and the weights are

In my experience, the accuracy gap between open-source and proprietary is negligible now and open-source is cheaper, faster, more customizable & sustainable for companies! No excuse anymore not to become an AI builder based on open-source AI (vs outsourcing to APIs)!

🎤 Excited to deliver an oral presentation on our paper, "Discovering Environments with XRM", tomorrow (Thursday) at #ICML2024! - What: #cross_risk_minimization #XRM - When: 10:45 - 11:00 CEST - Where: Oral 5E on OOD, rooms Straus 1-3 - ArXiv and Code: arxiv.org/abs/2309.16748

PSA: If you don't want your X posts used to train Grok, you now have to explicitly opt-out. Go to x.com/settings/grok_… and uncheck the box. If link doesn't work, go to Settings->Privacy and Safety->Grok

A big thanks to the #WiMLSymposium organizers! Your work is amazing!! #WiML #ICML Caroline Weis Mandana Samiei Maryleen Amaizu Irene Ballester Laya Rafiee Elizabeth Healey

📢 Montreal AI Symposium is back! It's a great venue to present your work and make connections to Montreal-based experts in AI. Contributions are in the format of two pages abstract and the deadline is August 22nd. Call for contributions: ow.ly/Uc9P50SKqcb

Llama 3.1 shows synthetic data is key for developing frontier-level models. i.e. develop a robust data gen pipeline and let GPUs go brr. Part of what Saba and team made is this pipeline for precise image editing, quite useful for the next open-source multimodal LLMs.

On behalf of MAIS board members, I am really excited to announce #MAIS 2024 Call for Paper. Please submit your work before August 22nd! MAIS is the largest AI Conference in Quebec, showcasing the work of researchers from industry and academia!

One thing I was very wrong about ~4yrs ago is how fundamental “synthetic data” in ML would be. “Obviously” due to data-proc-inequality, synth data should not help learning. The key flaw in this argument is: we do not use info-theoretically optimal learning methods in practice 1/

We are excited to launch the next cohort of the Cohere For AI Scholars Program — an opportunity to work with some of the best researchers in the world. 🎉 The Scholars Program is designed to change where, how, and by whom research is done. Apply by Aug 30! cohere-ai.ghost.io/cohere-for-ai-…

fyi: Meta's funding a few new grants for faculty to create evaluation benchmarks (deadline Sept 7). Could be of interest! #NLProc llama.meta.com/llm-evaluation…

Submission deadline for MAIS 2024 is this Thursday! published and unpublished works are welcome

The Call for Participation is out for this year's WiML workshop at NeurIPS! ✨ Submit by Sept 10. We also have travel funding available! Check out more details here: buff.ly/3WUvHgu

Consider submitting your work to our Safe Generative AI Workshop at NeurIPS 2024!

We have extended the deadline until Sep. 1st. MAIS is non-arcival, you can consider submitting your current or previous work to publicize it and recieve feedback. For more details, see call for contributions: montrealaisymposium.wordpress.com/call-for-contr…

Reminder that MAIS deadline is on Sept. 1st. CMT submission link: cmt3.research.microsoft.com/MAIS2024/ Also please consider signing up as a reviewer: forms.gle/NDGu1DjzhjQZck…

Diffusion is the rising tide that eventually submerges all frequencies, high and low 🌊 Diffusion is the gradual decomposition into feature scales, fine and coarse 🗼 Diffusion is just spectral autoregression 🤷🌈

As we move towards more powerful AI, it becomes urgent to better understand the risks in a mathematically rigorous and quantifiable way and use that knowledge to mitigate them. More in my latest blog entry where I describe our recent paper on that topic. yoshuabengio.org/2024/08/29/bou…

I’m excited to share that I’m serving as the Mentorship and Networking Chair at WiML in NeurIPS’24! We’d love to know which topics you’re most interested in discussing with our mentors. Please fill out this form: tinyurl.com/bdf5269p or write your suggestions below!