MAD in cartoons! Great depiction by CLOUD SCIENCE

When AI is trained on AI-generated content, it's basically data inbreeding. trib.al/lPe33yy

📢📢--OptimaLab Update-- 📢📢 It is as simple as this: To the best of our knowledge, this is the first paper on proving Nesterov's acceleration on deep MLPs (i.e., non-convex) under mild assumptions. Kudos to Jasper Liao (jasperliao.github.io) See: arxiv.org/pdf/2306.08109…

Recent work by Samuel "curry-howard fanboi" Ainsworth empirically evidences low loss linear paths between networks modulo permutations. In our work with amazing collaborators Goujaud Baptiste , Gauthier Gidel , @AymericD4 , we give theoretical insights on this phenomenon! arxiv.org/pdf/2310.19103…

I am happy to announce that our paper "self-consuming generative models go MAD" has been accepted as poster in the ICLR2024!

Excited that our paper "Pedagogical Alignment of LLMs" is accepted at EMNLP'24 findings 🎉 Thanks to all authors - Kangqi Ni, Sapana Chaudhary, Richard Baraniuk Read here: arxiv.org/abs/2402.05000

We have already achieved self-improvement in diffusion models, via its own synthetic data! Not that far away definitely!

🤖 From this week's issue: A research summary about self-improving diffusion models using synthetic data. buff.ly/4b66837

همین.