Shreshth Malik @ICML

AI4Science | Diffractive Labs | ML PhD @OATML_Oxford @aims_oxford

[Shreshth Malik @ICML @ShreshthMalik]

Thanks to co-authors @tiar_nan, Ben Colmey, Steve Roberts, @yaringal and Paul Midgley.
A great collab between @OATML_Oxford and @cu_mat.

Paper: https://t.co/tQbvKEbfdH
Code: https://t.co/x4CifIa4Hl full github release coming soon! (5/5)

[Shreshth Malik @ICML @ShreshthMalik]

Our work on hybrid physics-ML simulators for electron diffraction is now out in @NatureComms! ⚛️

A differentiable, GPU-accelerated, open-source package for dynamical structural refinement.🔬

Thread 🧵(1/5)

[Shreshth Malik @ICML @ShreshthMalik]

Result: fast, high-fidelity, interpretable refinement on synthetic and experimental data, with room to extend to other modalities (e.g. cryo-EM).

The codebase is hackable and enables fundamental research into electron diffraction (e.g. absorption https://t.co/v2ZNAVUcgN). (4/5)

[Charles 🎉 Frye @charles_irl]

LLMs are a "new material" for building computer applications. Fitting that they should also be used to discover novel physical materials!

Just one of many excellent papers in this year's ICML sponsored by @modal. By @ShreshthMalik of @OATML_Oxford et al.

https://t.co/lZk4Otekmr

[Yarin @yaringal]

🚀 Proud to share our new paper: MADE, a benchmark for closed-loop materials discovery🧪🤖

Discovery isn’t static: it’s adaptive, nonstationary & constrained

MADE tests LLM agents etc under feedback loops, OOD shifts & real-world constraints: pushing beyond regurgitating facts

[Shreshth Malik @ICML @ShreshthMalik]

Looking forward, envs like MADE enable evals and training of AI Scientists grounded in real-world discovery.

preprint: https://t.co/qfBsHl4E0X
Code: https://t.co/6BpGDFc8Lr

Thanks to co-authors at @OATML_Oxford @ImpMaterials and Diffractive Labs. And to @modal for compute!

[Shreshth Malik @ICML @ShreshthMalik]

Pleased to share that "MADE: Benchmark Environments for Closed-Loop Materials Discovery" was accepted at #ICML2026! 🇰🇷

We introduce a modular, open-source framework for evaluating AI scientists and algorithms at end-to-end, closed-loop materials discovery. 🤖

Thread 🧵(1/5)

[Shreshth Malik @ICML @ShreshthMalik]

We use Acceleration (AF) and Enhancement (EF) factors to quantify discovery efficiency relative to baselines.

We find that the uplift from using LLM agents and adaptive planning increases as chemical search spaces scale, and where surrogates get less reliable. 📈 (4/5)

[Yarin @yaringal]

Reminder - PhD applications for OATML are now open

The first funding deadline is December 2 - candidates interested in developing Bayesian deep learning methodology, applications of ML, AI security, and understanding ML methodology are encouraged to apply

More info: https://t.co/g99j1So31R

[Yarin @yaringal]

I'm at @ICCVConference and will be giving a keynote talk tomorrow at the @SaFeMMAI workshop at 10:55 (room 308 B)

Drop by to say hi!

I'm also looking for PhD students and postdocs to join @OATML_Oxford - happy to chat during the conference. Email: [email protected]

[Lukas Aichberger @aichberger]

⚠️Beware: Your AI assistant could be hijacked just by encountering a malicious image online!

Our latest research exposes critical security risks in AI assistants. An attacker can hijack them by simply posting an image on social media and waiting for it to be captured. [1/6] 🧵

[Jake Smith @crying_EM]

After thousands of microscope hours and several years of painstaking optimisation, @glynnca93 and I are very excited to share our latest manuscript. We outline our approach to investigate the molecular organisation within hippocampus tissue using cryo-ET.

[Sebastian Farquhar @seb_far]

Is your LLM hallucinating? 👻

Our @Nature paper shows how to detect when an LLM is making things up.

A 'confabulating' LLM answers with inconsistent meanings when re-asked the same question. We use this to estimate uncertainty and detect confabulations.

Learn more 🧵👇 1/

[Yarin @yaringal]

We're looking for 2 postdocs to join @OATML_Oxford, working on Foundational AI Safety. You’ll lead and contribute to projects aimed at developing principled and practical safe AI methods which would be used in real systems eg in the medical domain. Closing date 12th June

Links>>

[Jan Brauner @JanMBrauner]

Out in Science today:
In our paper, we describe extreme AI risks and concrete actions to manage them, including tech R&D and governance.
“For AI to be a boon, we must reorient; pushing AI capabilities alone is not enough.”

[OATML_Oxford @OATML_Oxford]

We’re organising a virtual open day!
📅 4-5:30 PM GMT on 22 Nov 2023

Anyone interested in joining @OATML_Oxford 🥣🌾for a PhD is welcome! Come to meet OATMLers, get a feel for our group, and ask any questions you may have!

✏️ Sign Up: https://t.co/zIzQpJNSxu