🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I'm happy to say that SCLD has been accepted at ICLR25!

Tweeting again about sampling, my favourite 2024 Monte Carlo paper is arxiv.org/abs/2307.01050 by F. Vargas, Shreyas Padhy, D. Blessing & N. Nüsken: . Propose a "simple" loss to learn the drift you need to add to Langevin to follow a fixed probability path.

Neat paper on classifier-free guidance by K. Pavasovic, J. Verbeek, Giulio Biroli & M. Mezard: arxiv.org/abs/2502.07849…

Working on sampling and seeking neural network ansatz? Longing for simulation-free* training approaches? – we review neural samplers and present a “failed” attempt towards it with pitfalls and promises! Joint work with Jiajun He (co-lead), Francisco Vargas … 🧵1/n

We have a new cool preprint with Cheuk Kit Lee. We developed a sweet Sequential Monte Carlo algorithm for unbiased samples from a tempered distribution p(x0)p(y|x0)^α and applied it to discrete diffusion for text arxiv.org/abs/2502.06079. Huge thanks to Francisco Vargas Michael Albergo

Amazing work by Jiajun He that generalises our work with Paul Jeha . Congrats!!

Intern brought in donuts on a Friday. People think that stuff is cringe but it’s not. I’m eating a Boston cream and the kid’s getting a return offer

Nice to see heartbeat scheduling in OxCaml Parallel library. Eliminates the need for granularity control in conjunction with stack local allocation of closures.

🧵(1/6) Delighted to share our ICML Conference 2025 spotlight paper: the Feynman-Kac Correctors (FKCs) in Diffusion Picture this: it’s inference time and we want to generate new samples from our diffusion model. But we don’t want to just copy the training data – we may want to sample