Log inSign up
Jason Yim
338 posts
user avatar
Jason Yim
@json_yim
Past: @Xaira_Thera, @MIT_CSAIL PhD, @GoogleDeepMind. Interests: generative models, LLMs, science.
Cambridge, MA
jasonkyuyim.com
Joined September 2017
370
Following
1,841
Followers
  • Pinned
    user avatar
    Jason Yim
    @json_yim
    Feb 25, 2024
    Combining discrete and continuous data is an important capability for generative models. To address this for protein design, we introduce Multiflow, a generative model for structure and sequence generation. Preprint: arxiv.org/abs/2402.04997 Code: github.com/jasonkyuyim/mu… 1/8
    GIF
    75K
  • user avatar
    Jason Yim
    @json_yim
    Oct 8, 2024
    Does this set precedence for AlphaFold to win the Nobel prize?
    user avatar
    The Nobel Prize
    @NobelPrize
    Oct 8, 2024
    BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”
    106K
  • user avatar
    Jason Yim
    @json_yim
    Oct 10, 2023
    Sharing an early preprint of my Microsoft AI4Science summer internship project. We developed SE(3) flow matching for protein backbone generation. Compared to SE(3) diffusion, we find our method achieves higher designability, faster sampling, with a way simpler implementation. 1/8
    GIF
    95K
  • user avatar
    Jason Yim
    @json_yim
    Apr 6, 2024
    Our review on diffusion models for protein structures and docking is out @HannesStaerk @GabriCorso Bowen Jing @BarzilayRegina Tommi Jaakkola. This summarize the advancements up to 2023. There have been lots of exciting new works since!
    Diffusion models in protein structure and docking
    From wires.onlinelibrary.wiley.com
    39K
  • user avatar
    Jason Yim
    @json_yim
    Feb 2, 2024
    Excellent review on how generative AI is transforming de novo protein design. I highly recommend.
    De novo protein design—From new structures to programmable functions
    From cell.com
    22K
  • user avatar
    Jason Yim
    @json_yim
    Apr 22, 2024
    My prediction for the next bio/ML trend at NeurIPS. DPO and RLHF for protein design. Protein language models in particular. 😉
    56K
  • user avatar
    Jason Yim
    @json_yim
    Oct 9, 2024
    Amazing choices. I had the chance to work on AlphaFold2-multimer at DeepMind then collaborate with Baker’s lab during my PhD. The talent and love for science at both places is phenomenal. Couldn’t be happier for the great scientists and engineers that I know made this happen.
    user avatar
    The Nobel Prize
    @NobelPrize
    Oct 9, 2024
    BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Chemistry with one half to David Baker “for computational protein design” and the other half jointly to Demis Hassabis and John M. Jumper “for protein structure prediction.”
    8.8K
  • user avatar
    Jason Yim
    @json_yim
    May 2, 2024
    Accepted to #ICML. Why do ICLR and ICML have to both be in Vienna 🙃
    user avatar
    Jason Yim
    @json_yim
    Feb 25, 2024
    Combining discrete and continuous data is an important capability for generative models. To address this for protein design, we introduce Multiflow, a generative model for structure and sequence generation. Preprint: arxiv.org/abs/2402.04997 Code: github.com/jasonkyuyim/mu… 1/8
    GIF
    15K
  • user avatar
    Jason Yim
    @json_yim
    Jul 22, 2024
    Our latest work extending FrameFlow for motif-scaffolding was published in TMLR. openreview.net/forum?id=fa1ne… Motif-scaffolding code (apologies for the delay): github.com/microsoft/prot…. **TL;DR**: diverse designable scaffolds, strong simple model for motif-scaffolding. (1/7)
    GIF
    7.8K
  • user avatar
    Jason Yim
    @json_yim
    Nov 4, 2023
    Replying to @json_yim @AndrewC_ML and 9 others
    We've released source code with MIT license
    GitHub - microsoft/protein-frame-flow: Fast protein backbone generation with SE(3) flow matching.
    From github.com
    15K
  • user avatar
    Jason Yim
    @json_yim
    Jul 15, 2023
    To emphasize this, we made RFdiffusion free and open source specifically so there is no paywall to advance protein design. Please do not pay money for these services.
    user avatar
    Helen Eisenach O'Brien
    @HelenEisenach
    Jul 15, 2023
    Just a reminder that RFdiffusion is free and publicly available, and thanks to @sokrypton, accessible through Google Colab! He's done an incredible job making it user friendly - no need to shell out the big $$$ colab.research.google.com/github/sokrypt…
    8.2K
  • user avatar
    Jason Yim
    @json_yim
    Sep 21, 2023
    NeurIPS reject justification: "The paper's heavy reliance on a black-box neural network to compute the fitness score undermines the reliability and interpretability of the results." Meanwhile other protein engineering papers with black box oracles gets in...
    16K
  • user avatar
    Jason Yim
    @json_yim
    Jan 9, 2024
    Excited to be organizing this ICLR workshop. Please consider submitting!
    user avatar
    GEMBio Workshop
    @gembioworkshop
    Jan 9, 2024
    Announcing the Generative and Experimental Perspectives for Biomolecular Design workshop at #iclr2024! We hope to bring together researchers in ML and experimental biology to accelerate progress on real-world applications. Website: gembio.ai Paper deadline: Feb 3
    6.9K
  • user avatar
    Jason Yim
    @json_yim
    Aug 19, 2024
    There are two types of researchers in this world. Diffusion: data is t=0 and noise is t=1. Flows: noise is t=0 and data is t=1. It's so much fun to read both types of papers and flip flopping between the two conventions! /s
    6.3K

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms of Service|Privacy Policy|Cookie Policy|Accessibility|Ads info|© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up