Kevin K. Yang 楊凱筌 (@KevinKaichuang) / X

Kevin K. Yang 楊凱筌

11.7K posts

Kevin K. Yang 楊凱筌

@KevinKaichuang

Principal Researcher in BioML @MSFTResearch (@MSRNE). He/him/他. 🇹🇼

Massachusetts, USA

Joined November 2016

Pinned
Kevin K. Yang 楊凱筌
@KevinKaichuang
May 17, 2019
We (@ZvxyWu, @francesarnold) wrote a review about machine learning for protein engineering, but there's so many new papers that we can't update it quickly enough. So Zach and I made a public list of papers covering machine learning on proteins. github.com/yangkky/Machin…
Kevin K. Yang 楊凱筌
@KevinKaichuang
Feb 19, 2021
What if I flip a coin 1e6 times, throw out half the heads, train an unbiased estimator on the remaining data, and then use it to decide who has to stay in jail pretrial?
Kevin K. Yang 楊凱筌
@KevinKaichuang
May 2, 2025
All my icml submissions got rejected, but we did have a different kind of success this week
122K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Jan 6, 2024
Idea for a dark academia novel: billionaire gives a guy fired from MIT for sleeping with a mentee millions to start a new lab, then gets Harvard's president fired for "plagiarism," then discovers his wife did the same thing so he's gonna investigate the rest of the MIT faculty
307K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Nov 18, 2023
msft gave openai $10B and all our gpus and openai is still firing people over Google meet
305K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Mar 28, 2024
And then they all got fired
Science Magazine
@ScienceMagazine
Mar 26, 2024
In a 2023 Science study, @AIatMeta researchers showed the power of a large language model—#ESMFold—to enable protein structure prediction and analysis. Using ESMfold, they generated a database of over 600 million metagenomic proteins. scim.ag/6mM #ScienceMagArchives
540K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Dec 14, 2024
Note: most MIT professors I know are honest and morally upright
99K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Sep 26, 2020
By having 0 papers of 0 submissions accepted to neurips, I've achieved the elusive acceptance rate of NaN.
Kevin K. Yang 楊凱筌
@KevinKaichuang
Feb 15, 2022
I (working at Microsoft) asked somebody from Facebook for data from a paper and got a Google drive link. Is this the singularity?
Kevin K. Yang 楊凱筌
@KevinKaichuang
Oct 3, 2022
We present Folding Diffusion: a diffusion model for protein structure inspired by physical protein folding. Let by @Kevin_E_Wu during his internship, with @alexijielu @vdbergrianne @james_y_zou @avapamini Code: github.com/microsoft/fold… Preprint: arxiv.org/abs/2209.15611
00:00
Kevin K. Yang 楊凱筌
@KevinKaichuang
Feb 8, 2024
We did 370 experiments to discover that protein language models primarily learn structure and won't scale for protein function prediction. We need new pretraining tasks! Work led by @francescazfl with @avapamini @yisongyue @alexijielu See Alex's thread + the paper for more!
Alex Lu
@alexijielu
Feb 8, 2024
Announcing our preprint understanding transfer learning for protein language models (PLMs), led by former MSRNE intern @francescazfl, with @KevinKaichuang @avapamini @yisongyue Key takeaway: PLMs do not scale for anything except structure! 🧵👇 biorxiv.org/content/10.110…
150K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Jan 10, 2023
You grow up thinking professors are smart, and then you see them donating to a sexual harasser with their real names attached
99K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Aug 9, 2023
Sad news as meta lays off their protein machine learning team. For a while, it was a valuable combination of scientific curiosity, openness, and industrial-scale engineering.
Meta disbands protein-folding team in shift towards commercial AI
From ft.com
173K
Kevin K. Yang 楊凱筌
@KevinKaichuang
Sep 13, 2023
EvoDiff combines evolutionary-scale data with diffusion models for controllable protein sequence generation. In addition to generating plausible proteins, we can scaffold structural motifs in sequence space! Preprint: biorxiv.org/cgi/content/sh… Code: github.com/microsoft/evod…
GIF
173K