Andrés M Bran
@drecmb
Followers
688
Following
36K
Media
50
Statuses
746
PhD student @SchwallerGroup @EPFL | Previously intern @mpiMathSci. I like to travel sometimes! 🇨🇴
Medellín, Colombia
Joined April 2016
RT @JeffGuo__: Generate property-optimized small molecules with 𝘴𝘵𝘦𝘦𝘳𝘢𝘣𝘭𝘦 𝘢𝘯𝘥 𝘨𝘳𝘢𝘯𝘶𝘭𝘢𝘳 synthesizability control - allowing complete user-fl….
0
19
0
@d_armstr is presenting Tango*: starting material-constrained retrosynthesis .
Happy to present a poster of my work on starting material constrained synthesis planning, Tango*, in the AI4Materials workshop at #ICLR2025 in 🇸🇬 Thanks to collaborators @JeffGuo__ and @ZJoncev and my PI.@pschwllr!.
0
0
2
@6ojaHa will present gollum: or how to beat every benchmark you know of, with GP-optimized LLMs 🫣.
How it feels to bash your own previous work in a new paper💀 Anywaaays, hello from #ICLR2025 in Singapore! 🇸🇬 .I’m presenting GOLLuM: GP-optimised LLMs at the World Models Workshop on Monday. Let's discuss LLMs, finetuning, GPs, BO, life, pain, PhD👀.🔗
1
0
3
@JunwuChen25 will present on RL for 3D crystal diffusion generation
Thrilled to present our new work on Reinforcement Learning for 3D Crystal Diffusion Generation in #AI4Mat at #ICLR2025 ! We show that RL can optimize diffusion models for goal-directed crystal generation. Thanks to @pschwllr and @SchwallerGroup !.
1
0
2
Also don't miss the other members of our group who'll be around this week 🤩. @RebeccaNeeser with her work on fragment-based drug design
Hello from Singapore 🇸🇬! Thrilled to be at #ICLR2025 presenting our work on fragment-based drug discovery 🧩. We go beyond virtual screening with a generative, structure-aware approach. 📃 🔗 A thread 🧵👇.
1
0
2
I'll be presenting at the #AI4Mat and #llmreasoning workshops, our work on reasoning in LLMs, with @TheoNeukomm . 🧵: 📎: 🧑💻:
LLMs are pretty bad at writing molecules, but quite good at analyzing mols and reactions!.In our new work we use LLMs+search in chemical tasks, unlocking steerable synth. planning and mechanism prediction 🌟. 1/.@TheoNeukomm @d_armstr @ZJoncev @pschwllr .
1
2
3
This week we're in #ICLR2025 in Singapore 🇸🇬!!. Do reach out to discuss the future of #AI and #Chemistry 🥂 llms, agents, new tools for chemists and new ways of doing chemistry 🚀.
2
2
12
RT @RebeccaNeeser: Hello from Singapore 🇸🇬! Thrilled to be at #ICLR2025 presenting our work on fragment-based drug discovery 🧩. We go beyon….
0
19
0
RT @JunwuChen25: Thrilled to present our new work on Reinforcement Learning for 3D Crystal Diffusion Generation in #AI4Mat at #ICLR2025 !….
0
5
0
RT @JunwuChen25: AdsMT is finally out in @NatureComms 🥳. It is designed for rapid prediction of global minimum adsorption energy (GMAE) fro….
0
8
0
Many other problems in chemistry can be tackled like this! From the fundamental stuff to larger scales, LLMs might already be enough for searching thru complex chemical spaces. Kudos to all the authors @TheoNeukomm @d_armstr @ZJoncev @pschwllr from @SchwallerGroup @EPFL_en !.
1
0
4
We also tackled mechanism determination, with also great results from @TheoNeukomm! LLMs are also quite good at guiding search for mechanisms. The cooler thing is that, as with synthesis, arbitrary inputs can be given in, to guide the search, all in words!
1
0
3
We made a small benchmark for this task **steerable synthesis planning**, with several targets and prompts, and the results are quite impresive!. Larger models are quite good at the task 🚀, with newer versions improving rapidly. A small model like gpt-4o-mini is not as good tho
1
0
5
In synthesis planning, we input mol. and the output are some routes, with no restrictions. But what if I want to specify details? - see img.In our work, users can specify a query, and during search the LLM selects only aligned solutions - working as a zero-shot value function!
1
0
4
Many tasks in chemistry can be tackled through search, like synthesis planning, mechanism determination, molecular design, etc. Current methods rely on designed or learned value functions, but what if LLMs already know all of this? and how can we exploit it?.
1
0
5