Joel Shor
@joelshor.bsky.social
AI for drug discovery, AI for social good, retired cultural truffle pig
🚀 Our work with @Google Research is now featured on the Google AI Blog!
🧬 NucleoBench — the first large-scale benchmark for nucleic acid sequence design
✨ AdaBeam — a new optimizer that sets state-of-the-art on 11/16 tasks
Full story 👉 research.google/blog/smarter...
🧬 NucleoBench — the first large-scale benchmark for nucleic acid sequence design
✨ AdaBeam — a new optimizer that sets state-of-the-art on 11/16 tasks
Full story 👉 research.google/blog/smarter...
Smarter nucleic acid design with NucleoBench and AdaBeam
research.google
September 11, 2025 at 5:59 PM
🚀 Our work with @Google Research is now featured on the Google AI Blog!
🧬 NucleoBench — the first large-scale benchmark for nucleic acid sequence design
✨ AdaBeam — a new optimizer that sets state-of-the-art on 11/16 tasks
Full story 👉 research.google/blog/smarter...
🧬 NucleoBench — the first large-scale benchmark for nucleic acid sequence design
✨ AdaBeam — a new optimizer that sets state-of-the-art on 11/16 tasks
Full story 👉 research.google/blog/smarter...
Thrilled our work on nucleic acid design with Google was accepted to the ICML 2025 GenBio workshop! 🧬:
1. NucleoBench: A large-scale benchmark for nucleic acid sequence design
2. AdaBeam: A new optimizer that outperforms others on 11/16 tasks.
📖 Paper: lnkd.in/e-_q-2fP
💻 Code: lnkd.in/ewM8ucxh
1. NucleoBench: A large-scale benchmark for nucleic acid sequence design
2. AdaBeam: A new optimizer that outperforms others on 11/16 tasks.
📖 Paper: lnkd.in/e-_q-2fP
💻 Code: lnkd.in/ewM8ucxh
NucleoBench: A Large-Scale Benchmark of Neural Nucleic Acid Design Algorithms
One outstanding open problem with high therapeutic value is how to design nucleic acid sequences with specific properties. Even just the 5' UTR sequence admits 2*10^120 possibilities, making exhaustiv...
www.biorxiv.org
June 26, 2025 at 6:05 PM
Thrilled our work on nucleic acid design with Google was accepted to the ICML 2025 GenBio workshop! 🧬:
1. NucleoBench: A large-scale benchmark for nucleic acid sequence design
2. AdaBeam: A new optimizer that outperforms others on 11/16 tasks.
📖 Paper: lnkd.in/e-_q-2fP
💻 Code: lnkd.in/ewM8ucxh
1. NucleoBench: A large-scale benchmark for nucleic acid sequence design
2. AdaBeam: A new optimizer that outperforms others on 11/16 tasks.
📖 Paper: lnkd.in/e-_q-2fP
💻 Code: lnkd.in/ewM8ucxh
Excited to be giving a virtual talk on a nucleic acid design collaboration tomorrow. Organizers are Nucleate, a student-led biotech group. Work was done by my techbio Consulting company, Move37 Labs, and collaborators at MIT and Google. Sign-up for the talk here: lu.ma/aixbio-tech-...
February 18, 2025 at 6:22 PM
Excited to be giving a virtual talk on a nucleic acid design collaboration tomorrow. Organizers are Nucleate, a student-led biotech group. Work was done by my techbio Consulting company, Move37 Labs, and collaborators at MIT and Google. Sign-up for the talk here: lu.ma/aixbio-tech-...
🎉 New research publication, with the brilliant Carson McNeil!
We have a surprising result: we can predict the performance of a medical ML model on a new dataset, without having labels on that dataset!
Read more: rdcu.be/d0Owc
Why this matters:
(see below)
We have a surprising result: we can predict the performance of a medical ML model on a new dataset, without having labels on that dataset!
Read more: rdcu.be/d0Owc
Why this matters:
(see below)
Predicting the generalization of computer aided detection (CADe) models for colonoscopy
rdcu.be
November 20, 2024 at 4:12 PM
🎉 New research publication, with the brilliant Carson McNeil!
We have a surprising result: we can predict the performance of a medical ML model on a new dataset, without having labels on that dataset!
Read more: rdcu.be/d0Owc
Why this matters:
(see below)
We have a surprising result: we can predict the performance of a medical ML model on a new dataset, without having labels on that dataset!
Read more: rdcu.be/d0Owc
Why this matters:
(see below)