Log inSign up
Image
user avatar
Simon Kohl
@saakohl
Founder at Latent Labs.
simonkohl.com
Joined December 2016
445
Following
2,991
Followers
  • Pinned
    user avatar
    Simon Kohl
    @saakohl
    Mar 23
    Today we're launching Latent-Y: the world's first autonomous agent for drug design, lab-validated end to end. Give it a research goal. Latent-Y reasons, designs, iterates, and delivers lab-ready antibodies, autonomously or collaboratively, with the biological reasoning of a PhD
    Image
    00:00
    120K
  • user avatar
    Simon Kohl
    @saakohl
    Sep 14, 2018
    Our paper `A Probabilistic U-Net for Segmentation of Ambiguous Images' was accepted at #NIPS2018 as a spotlight presentation! A re-implementation of the code is now available at github.com/SimonKohl/prob…. Paper arxiv.org/abs/1806.05034 by @DeepMindAI and @mic_dkfz.
    Image
    GIF
  • user avatar
    Simon Kohl
    @saakohl
    Feb 13, 2025
    @Latent_Labs comes out of stealth today with $50M funding. Our goal? To push the frontiers of generative biology, giving partners instant access to tools capable of accelerating drug design. Every biotech or pharma company searching for the best therapeutic molecules understands
    Image
    00:00
    46K
  • user avatar
    Simon Kohl
    @saakohl
    Jul 12, 2022
    Come work with us as a research scientist on ML for structural & synthetic biology! Apply here 👉boards.greenhouse.io/deepmind/jobs/…
  • user avatar
    Simon Kohl
    @saakohl
    Feb 28, 2017
    `Adversarial Networks for the Detection of Aggressive Prostate Cancer' - arxiv.org/abs/1702.08014 @soumithchintala @goodfellow_ian
    Image
  • user avatar
    Simon Kohl
    @saakohl
    Dec 7, 2020
    We published 'nnU-net' in @naturemethods today! It scores in the top ranks of 23 public biomedical datasets - w/o any manual interventions! What's the secret sauce? Rigorous development on many diverse datasets + a healthy dose of domain knowledge. rdcu.be/cbOJJ
  • user avatar
    Simon Kohl
    @saakohl
    Dec 5, 2018
    Come out tomorrow morning at #NeurIPS to hear about our work on Probabilistic Image Segmentation: spotlight: Room 220 E, 10:35-10:40am poster: #127, Room 517 AB (upstairs!), 10:45-12:45am paper: arxiv.org/abs/1806.05034 github: github.com/SimonKohl/prob… Looking forward! :)
    Image
  • user avatar
    Simon Kohl
    @saakohl
    May 31, 2019
    Medical images can exhibit ambiguities on multiple scales & locations often varying independently. We propose a hierarchical generative model to capture such variations in segmentations & show much improved sample fidelity and fit with the GT distribution:
    arXiv logo
    arxiv.org
    A Hierarchical Probabilistic U-Net for Modeling Multi-Scale Ambiguities
    Medical imaging only indirectly measures the molecular identity of the tissue within each voxel, which often produces only ambiguous image evidence for target measures of interest, like semantic...
  • user avatar
    Simon Kohl
    @saakohl
    Jul 15, 2021
    We're sharing the AlphaFold code on github and describe its details in @Nature today! Happy folding! :)
    user avatar
    Demis Hassabis
    @demishassabis
    Jul 15, 2021
    Last year we presented #AlphaFold v2 which predicts 3D structures of proteins down to atomic accuracy. Today we’re proud to share the methods in @Nature w/open source code. Excited to see the research this enables. More very soon! bit.ly/alphafoldmetho… bit.ly/alphafoldgithub
    Image
  • user avatar
    Simon Kohl
    @saakohl
    Nov 30, 2020
    Incredibly excited to share that #AlphaFold2 predicts protein structures to unparalleled accuracy in #CASP14! So proud and humbled to be part of the team! :)
    user avatar
    Google DeepMind
    @GoogleDeepMind
    Nov 30, 2020
    In a major scientific breakthrough, the latest version of #AlphaFold has been recognised as a solution to one of biology's grand challenges - the “protein folding problem”. It was validated today at #CASP14, the biennial Critical Assessment of protein Structure Prediction (1/3)
    Image
    GIF
  • user avatar
    Simon Kohl
    @saakohl
    Jun 14, 2018
    Internship work at @DeepMindAI on conditional density models over segmentations: `A Probabilistic U-Net for Segmentation of Ambiguous Images' (arxiv.org/abs/1806.05034). Joint w. @ber24, @clemenslm, @JeffreyDeFauw, Joe Ledsam, @maierhein, @arkitus, @DeepSpiker and Olaf Ronneberger
    arXiv logo
    arxiv.org
    A Probabilistic U-Net for Segmentation of Ambiguous Images
    Many real-world vision problems suffer from inherent ambiguities. In clinical applications for example, it might not be clear from a CT scan alone which particular region is cancer tissue....
  • user avatar
    Simon Kohl
    @saakohl
    Oct 9, 2024
    Goosebumps!
    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.”
    Image
    3.6K
  • user avatar
    Simon Kohl
    @saakohl
    Jul 14, 2020
    New work w/ colleagues @googlehealth & @DeepMind : `Contrastive Training for Improved Out-of-Distribution Detection' Contrastive Training + Label Smoothing surpasses previous OOD Detectors -no extra Data or OOD examples required! Bonus: We measure how far OOD the test data is.
    user avatar
    Olaf Ronneberger
    @ORonneberger
    Jul 14, 2020
    (1/2) Our new paper "Contrastive Training for Improved Out-of-Distribution Detection" arxiv.org/abs/2007.05566 with @jimwinkens, @BunelR, @abzz4ssj, Robert Stanforth, @vivnat, @joe_ledsam, @patmacwilliams, @pushmeet, @alan_karthi, @saakohl, @TaylanCemgilML, @arkitus
    Image
  • user avatar
    Simon Kohl
    @saakohl
    Dec 12, 2024
    I had an incredible time this week joining John and Demis together with the rest of the AlphaFold 2 team in Stockholm for the Nobel Prize festivities. In his Nobel lecture, John recounted the history behind AlphaFold 2's development including the the many breakthroughs ..
    Image
    18K