Log inSign up
user avatar
Nima Shoghi
@nimawrites
MTS @boltz_bio | Prev: @GeorgiaTech
London, United Kingdom
nima.sh
Joined February 2017
332
Following
221
Followers
RepliesMedia
  • Pinned
    user avatar
    Nima Shoghi
    @nimawrites
    Feb 3
    Excited to share STAR-MD: a scalable autoregressive diffusion model that generates stable, high-quality protein MD trajectories at microsecond timescales, where existing methods fail catastrophically. Accepted at ICLR 2026! Links and details in the thread 🧵👇 1/7
    Image
    00:00
  • user avatar
    Nima Shoghi
    @nimawrites
    May 7, 2024
    🎉Super excited to have presented our poster for "From Molecules to Materials: Pre-training Large Generalizable Models for Atomic Property Prediction" at ICLR 2024! Check out our webpage for our code, interactive visualization, and more! nima.sh/jmp/
    Image
  • user avatar
    Nima Shoghi
    @nimawrites
    Oct 30, 2023
    🎉 Excited to introduce Joint Multi-domain Pre-training (JMP) - our new pre-training method leveraging diverse chemical data at scale. JMP achieves SOTA performance on 34/40 evaluated tasks from QM9, rMD17, MD22, SPICE, MatBench, and QMOF. (1/5) arxiv.org/abs/2310.16802
    An overview of the Joint Multi-domain Pre-training (JMP) method. The left image demonstrates the joint pre-training setup, where a single model is simultaneously trained on set of diverse pre-training datasets using multi-task learning. The right image shows the fine-tuning process, where the pre-trained JMP backbone is equipped with new prediction heads and trained on downstream tasks.