Progress on Our BRAINIAC Initiative 🧠
Just submitted our NSF ACCESS allocation progress report, and I wanted to share some highlights! The BRAINIAC project (Broad‐scope Reasoning Artificial Intelligence for Nano-micro Identification, Assessment, and Categorization) is really taking off! 🌟
What we’ve accomplished 📊
- Our proof-of-concept is now published as a cover article in Molecular Systems Design & Engineering! 🎉
- Our SGNN model achieved 89% classification accuracy on sensor data - that’s 4x better than random baselines!
- We’ve successfully extracted and structured data from over 1,400 research papers using LLM-assisted text mining
BRAINIAC combines three powerful approaches:
- LLM-Assisted Text Mining to convert research papers into structured datasets 📚
- Spiking Graph Neural Networks that fuse global properties with molecular details 🔄
- High-Fidelity Simulations (DFT/AIMD) for mechanistic validation 🔬
The coolest part? We’re bridging the gap between isolated ML approaches and how human experts actually think and reason about materials!
Exciting News: Cover Article in Molecular Systems Design & Engineering! 🌟
I’m thrilled to share that our research on field-effect transistor chemical sensor design has been featured as the cover article in Molecular Systems Design & Engineering (Issue 5, 2025)! 🥳
What’s the research about? 🧪
Our team developed a unique spiking graph neural network (SGNN) architecture (brilliantly proposed by our PhD candidate Rodrigo Ferreira). This innovative approach effectively handles physicochemical cheminformatics by encoding information via two modalities (graph and spike) within field effect transistor chemical sensor knowledge graphs.
In simple terms: we created a smarter way to design chemical sensors using AI that thinks more like the human brain! 🧠
Real-world impact 🌎
As a result, we significantly enhanced the virtual screening of optimal sensor probes for the challenging detection of per- and polyfluoroalkyl substances (PFAS) - those troublesome “forever chemicals” that persist in our environment.
The paper has also been selected as part of MSDE’s “Recent HOT Articles” collection! 🔥
Want to read more?
Check out the full paper here: Expediting field-effect transistor chemical sensor design with neuromorphic spiking graph neural networks
Huge thanks to my amazing collaborators Rodrigo P. Ferreira, Fengxue Zhang, Haihui Pu, Claire Donnat, Yuxin Chen, and Junhong Chen! 👏
By the way this would be a very important milestone for me to realize the vision of BRAINIAC. Probably would submit some progress reports soon…
Big News: Our Research Published in Science Advances! 📣
I’m incredibly excited to announce that our latest work on multistage machine learning for catalyst discovery has been published in Science Advances! 🥂
What’s the breakthrough? ⚡
In this paper, we introduced a multistage machine learning framework that integrates three powerful techniques:
- Data mining 🔍
- Active learning 🔄
- Domain adaptation 🔀
This integrated approach helps us streamline the discovery of multimetallic catalysts - materials that are essential for clean energy technologies.
How does it work? 🤔
Our framework leverages different data modalities through specialized ML modules. This allows us to objectively navigate an enormous candidate space of possible catalyst combinations.
The result? We identified a promising catalyst with excellent performance in wet-lab testing and real commercial potential! 💰
The whole paper is intuitive but relatively extensive in labor. So recommended to read the full text.
Read the full paper 📝
Check out our paper: Leveraging Data Mining, Active Learning, and Domain Adaptation in a Multi-Stage, Machine Learning-Driven Approach for the Efficient Discovery of Advanced Acidic Oxygen Evolution Electrocatalysts
Preprint available on arXiv: arXiv:2407.04877
Huge thanks to my amazing collaborators especially my advisors Yuxin Chen, and Junhong Chen! 🙏
🎉 Exciting News! 🎉
I’m thrilled to announce that our project, BRAINIAC: An AI-Driven Framework for Accelerating Nano-Micro Materials Discovery and Device Innovation, has received a substantial allocation from NSF ACCESS — 1.5 million credits! This support will be instrumental in scaling up our efforts and pushing the boundaries of AI-driven materials discovery.
BRAINIAC has already garnered major institutional support, with awards from Argonne National Laboratory’s Nanoscale Materials and the 2024 AI+Science Research Initiative at University of Chicago’s Data Science Institute. These resources have enabled us to process 40,000+ publications using LLMs and conduct thousands of DFT simulations for critical applications like detecting PFAS in water.
Special thanks to Professor Junhong Chen and Professor Yuxin Chen for their ongoing support and guidance! I’m incredibly grateful for their mentorship, which makes all of this possible.
We’re also excited to share that our passionate PhD candidate, Rodrigo Pires Ferreira (rpferreira@uchicago.edu), is making fantastic progress on preliminary results for BRAINIAC. We plan to submit an initial manuscript by the end of 2024! 📈
With the ACCESS allocation, we’re empowered to explore new possibilities in AI and materials science, contributing to Uchicago’s leadership in AI research and paving the way for breakthrough discoveries.
Stay tuned for more updates as we continue this exciting journey! 🚀
Exciting News!
We are thrilled to announce that our review has been accepted by Chemical Society Reviews.
You can read our review online.
And there is the Uchicago PME news article online.
We published a new paper on ArXiv:
https://arxiv.org/abs/2407.04877
Onboarding in UChicago and ANL.