NSF Center for Computer Assisted Synthesis Creates Opportunities for Discovery Carnegie Mellon chemistry professors, Olexandr Isayev and Gabriel Gomes, are making breakthroughs as part of the NSF Center for Computer Assisted Synthesis.

📰C-CAS in the News 📰

This article highlights the work of C-CAS PIs @olexandr.bsky.social and @gabegomes.bsky.social as well as trainees Robert MacKnight, Nick Gao, and Liliana Gallegos.

www.cmu.edu/mcs/news-eve...

#chemistry #computerscience #nsf #chemsky

Photo of Jose smiling and holding two thumbs up. Text reads "Meet Jose Regio, PhD student at Carnegie Mellon University." CMU and C-CAS logos are also displayed.

Photo of Jose presenting a research poster. Text reads "Jose’s research focuses on designing visual interfaces for researchers in chemistry. His aim is to improve the usability of scientific tools and to make data easier to explore. By removing the barrier of extensive programming experience, this work gives experimental chemists access to computational tools and insights from big datasets."

Photo of Jose in front of a M. Pimatubo Crater Lake. Text reads "Jose was born in the Philippines and grew up in Connecticut. Outside of the lab, Jose is a volunteer at the Latino Community Center in Pittsburgh where he helps with math lessons and fostering teamwork for elementary schoolers. In his free time he likes to learn Tagalog, Spanish, and Mandarin."

Meet Jose Regio, a PhD student in the lab of @gabegomes.bsky.social at @cmuchemistry.bsky.social.

Have a question for Jose? Let us know in the replies.

#nsf #chemistry #ccas #meetascientist #computerscience #science #phd #stem #chemsky
1m

Will AI ever win its own Nobel? Some predict a prize-worthy science discovery soon Other researchers question whether autonomous AI scientists are possible or even desirable.

CMU researchers are pushing the boundaries of AI in science. Gabe Gomes and his team developed Coscientist, an AI system that plans and executes chemical reactions — accelerating discoveries that once took years to complete. His work was featured in Nature.
www.nature.com/articles/d41...

honestly, excellent summary!

CATNIP for the win! Read our newest work with the Gomes group- doi.org/10.1038/s415...

@gabegomes.bsky.social @alisonnarayan.bsky.social @aepaton.bsky.social

Rethinking chemical research in the age of large language models - Nature Computational Science This Perspective highlights the potential integrations of large language models (LLMs) in chemical research and provides guidance on the effective use of LLMs as research partners, noting the ethical and performance-based challenges that must be addressed moving forward.

A Perspective from @gabegomes.bsky.social and colleagues covers the opportunities for LLMs to advance chemical research and the challenges that must be overcome to effectively use LLMs as scientific partners. #chemsky

➡️ nature.com/articles/s43...

🔓 rdcu.be/etcqs

🧵(8/13)

not a gotcha/rhetorical question: where/what sector do you think the most interesting science is being done today, broadly speaking? i’m trying to understand something.

Pre-trained knowledge elevates large language models beyond traditional chemical reaction optimizers Modern optimization in experimental chemistry employs algorithmic search through black-box parameter spaces. Here we demonstrate that pre-trained knowledge in large language models (LLMs) fundamentall...

preprint: arxiv.org/abs/2509.00103 [5/5]

access Iron Mind at gomes.andrew.cmu.edu/iron-mind/.

excellent work led CMU Ph.D. students Robert MacKnight (ChemE) with Jose Regio (Chemistry), in collaboration with our colleagues Jeffrey Ethier and Luke Baldwin from AFRL. [4/5]

Iron Mind was developed to facilitate community engagement and set new benchmarks for optimization strategies and foundation models. it enables direct comparison of human, algorithmic, and LLM optimization campaigns on public leaderboards. [3/5]

our new preprint "Pre-trained knowledge elevates large language models beyond traditional chemical reaction optimizers" demonstrates that LLMs consistently match or exceed state-of-the-art BO across diverse chemical reaction datasets. [2/5]

new preprint, tl;dr:
• LLMs match or exceed SOTA strategies on chemical reaction optimizations.
• we built the Iron Mind platform and we hope that it can serve as a new benchmark for both reaction optimizers and foundation models.

a thread [1/5]

Team Provides a Roadmap for Large Language Models in Chemical Research - Mellon College of Science - Carnegie Mellon University In Nature Computational Science, Gomes and his coauthors offer a roadmap toward more strategic implementations of LLMs.

Gabe Gomes is rethinking how AI supports chemistry, taking the technology from providing passive answers to being an active lab partner that can help plan, test and accelerate discovery.
www.cmu.edu/mcs/news-eve...

Rethinking chemical research in the age of large language models - Nature Computational Science This Perspective highlights the potential integrations of large language models (LLMs) in chemical research and provides guidance on the effective use of LLMs as research partners, noting the ethical ...

A Perspective from @gabegomes.bsky.social and colleagues covers the opportunities for LLMs to advance chemical research and the challenges that must be overcome to effectively use LLMs as scientific partners.

www.nature.com/articles/s43...

🔓https://rdcu.be/etcqs

i made chatGPT (o3-pro + deep research) work for 63m on a task. i’ve never had it take any longer than 30 min on anything before this one. nice.

Enhancing Molecular Machine Learning with Quantum-Chemical Insight - Mellon College of Science - Carnegie Mellon University In Nature Machine Intelligence, Gabe Gomes, Daniil Boiko and collaborators propose a new type of molecular ML representation.

In Nature Machine Intelligence, Faculty member Gabe Gomes @gabegomes.bsky.social , PhD student Daniil Boiko, and their collaborators propose a new type of molecular ML representation that includes quantum-chemical interaction

www.cmu.edu/mcs/news-eve...

It was a great pleasure to host @groupquintana.bsky.social for @cmuscience.bsky.social seminar! RamO(N) was talking about linear sling methods in molecular similarity and fast clustering! #chemsky #compchem Selfie with @gabegomes.bsky.social

Deafheaven - Winona (Short Film) YouTube video by Deafheaven

and the short film they made is also a beautiful piece of art on its own: youtu.be/4cjJ-b-W_7A?...

Deafheaven - Incidental II (feat. Jae Matthews) (Official Audio) YouTube video by Deafheaven

stuff is all weird, but at least Deafheaven delivered a masterpiece with their latest album “Lonely People With Power”. definitely not everyone’s cup of tea (tw: heavy metal, harsh vocals), but it’s beautiful: youtu.be/XN7o7l5eehk?...

Chemists create matchmaking app for biocatalysis Machine learning model Catnip helps chemists find biocatalysts to try in synthesis

Excited to share this article from @cenmag.bsky.social! Such a nice highlight of the work we've been doing between @narayanlab.bsky.social and @gabegomes.bsky.social

cen.acs.org/acs-news/acs...

Chemists create matchmaking app for biocatalysis Machine learning model Catnip helps chemists find biocatalysts to try in synthesis

@narayanlab.bsky.social and @gabegomes.bsky.social designed Catnip--so named because it's "for the biocats"--to make it easier for newbs to use enzymatic reactions in organic synthesis. I had the privilege of hearing them talk about it at #ACSSpring2025 earlier this week!
cen.acs.org/acs-news/acs...

Chemists create matchmaking app for biocatalysis Machine learning model Catnip helps chemists find biocatalysts to try in synthesis

A team led by the @narayanlab.bsky.social and @gabegomes.bsky.social has developed an app to determine which biocatalysts are likely to work for a given molecule, making it easier for chemists to incorporate biocatalysts into organic synthesis.
@alisonnarayan.bsky.social

cen.acs.org/acs-news/acs...

Don't miss @gabegomes.bsky.social at ACS today at 4:30 PM PT!

More info about his talk here: acs.digitellinc.com/live/34/sess...

#chemsky #NSF #CCAS #chemistry #ACS

day 2 at the ACS Spring 2025: Daniil talks CATNIP (3:20 PM, Hall F Room 1). Robert (COMP Poster Session, 7 PM) will cover 'the Iron Mind project' and @gabegomes.bsky.social (4:30 PM, Room 25A) will give a glimpse of our efforts towards autonomous catalyst optimization with LLMs.

Check out our work on corrected product formulas (CPF) to lower the cost of #hamiltonian #simulation #quantumalgorithms @uoftcompsci.bsky.social @cqiqc-uoft.bsky.social @vectorinstitute.ai @chemuoft.bsky.social

the @gpggrp.bsky.social is at the ACS Spring 2025! come check out the works of Daniil Boiko and Rob MacKnight at the "ML + AI in Organic Chemistry" Symposium (Hall B-1, Room 4) today! extreme scaling of experimental chemical reactions via MS and an OS for autonomous comp chem!

hello San Diego!

People are rushing to apply AI to discover materials for applications in health, energy, aerospace and more. Are we the cool kids now?

- Lila => $200M
- OpenAI VP => AI+materials company
- Cusp => $30M
- RadicalAI automating materials discovery
- FutureHouse scientific agents

Despite being trained primarily on monophosphine ligands in Pd(0)/Pd(II) cycles, it demonstrates good transferability to diverse palladium complexes, including bidentate phosphines and dinuclear Pd(I) complexes. The speed of AIMNet2-Pd makes computational high-throughput catalyst screening feasible

The model can characterize entire catalytic cycles including geometry optimization, transition state searches, and energy calculations for each step of a catalytic process (oxidative addition, transmetalation, and reductive elimination etc).