Our new paper, IB-EDL, introduces a novel method to enhance trustworthiness in fine-tuned LLMs by regularizing Evidential Deep Learning with an information bottleneck for improved uncertainty estimation and calibration.excellent work by @yaweili-sandylaker.bsky.social, @davidruegamer.bsky.social

Slides for a general introduction to the use of Optimal Transport methods in learning, with an emphasis on diffusion models, flow matching, training 2 layers neural networks and deep transformers. speakerdeck.com/gpeyre/optim...

NeurIPS 2024 starts today and MCML researchers are represented with 26 accepted papers! 🎉

Here is an overview of the MCML papers at the conference: mcml.ai/news/2024-12...

Congrats to our researchers!

#NeurIPS #MachineLearning #AI @neuripsconf.bsky.social

If you reviewed for #ICLR, please make sure to read other reviewers' comments too and reflect on whether you may have missed something.

The paper will need to have a single decision; the point of this exercise is not just about addressing each reviewer's concerns individually.

It's tough to gain visibility as a young researcher, and it's job market season! Are you a theoretical computer science PhD/postdoc on the job market?

I don't have a crazy juge audience but I'll try to help a bit: fill this form, and I'll tweet your pitch and info!
docs.google.com/forms/d/e/1F...

PhD Opportunity to work with Christian Cabrera and me on Self Sustaining Software Systems.

Part of the Manchester-Cambridge CDT.

www.findaphd.com/phds/project...