MolSC: Leveraging Substituent Contributions to Enhance Fine-grained Molecular Understanding in LLMs
We propose MolSC, a method that leverages substituent contributions to enhance fine-grained molecular understanding in large language models.
Professional Summary
Alex Johnson is a Senior AI Research Scientist at Meta AI. His research
has been published in top conferences like NeurIPS and ICML, with over
10,000 citations. Alex is passionate about pushing the boundaries of AI
while ensuring ethical development.
Education
PhD Computer Science
2015-09-01
2019-06-30
Stanford University
MS Computer Science
2013-09-01
2015-05-31
Carnegie Mellon University
BS Computer Science
2009-09-01
2013-05-31
MIT
Interests
Large Language Models
Computer Vision
Reinforcement Learning
AI Ethics
Research
My research spans three interconnected areas: vision-language modeling, commonsense reasoning, and multi-modal chemical language modeling. I am particularly interested in data refinement and synthesis strategies that enable language models to better understand the physical and chemical world.
I am always open to collaboration — feel free to reach out.
Featured Publications
Bridging the Gap Between Molecule and Textual Descriptions via Substructure-aware Alignment
We propose a substructure-aware alignment approach to bridge the gap between molecular and textual representations.
Recent Publications
(2026).
MolSC: Leveraging Substituent Contributions to Enhance Fine-grained Molecular Understanding in LLMs.
Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026, Main).
(2026).
Enhancing Zero-shot Commonsense Reasoning by Integrating Visual Knowledge via Machine Imagination.
Under Review.
(2025).
Bridging the Gap Between Molecule and Textual Descriptions via Substructure-aware Alignment.
Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP 2025, Main).
(2024).
Moleco: Molecular Contrastive Learning with Chemical Language Models for Molecular Property Prediction.
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track (EMNLP 2024, Industry).
(2024).
MolTRES: Improving Chemical Language Representation Learning for Molecular Property Prediction.
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP 2024, Main).
Recent & Upcoming Talks
Create Beautiful Presentations with Markdown
Discover how to create stunning, interactive presentations using simple Markdown — no PowerPoint, Keynote, or vendor lock-in required.
Recent News
Trending
⚡️ Turn Jupyter Notebooks into Blog Posts
Publish your data science and research directly from Jupyter Notebooks. No screenshots required.
Trending
🎉 Easily create your own simple yet highly customizable blog
Take full control of your personal brand and privacy by migrating away from the big tech platforms!
🧠 Sharpen your thinking with a second brain
Create a personal knowledge base and share your knowledge with your peers.
📈 Communicate your results effectively with the best data visualizations
Use popular tools such as HuggingFace, Plotly, Mermaid, and data frames.
👩🏼🏫 Teach academic courses
Embed videos, podcasts, code, LaTeX math, and even test students!
✅ Manage your projects
Easily manage your projects - create ideation mind maps, Gantt charts, todo lists, and more!