Georgia Channing
I lead the AI for Science team at Hugging Face. I am also a researcher affiliated with Torr Vision Group at the University of Oxford. My LinkedIn says that I used to think about languages and now I think about science, and that's pretty much true. Over my academic career, I worked in many different scientific domains while broadly interested in representation learning, and now I do that professionally. I spend most of my time thinking about what high quality science is and how we can use ML to enable it.
georgia [at] huggingface [dot] co [no m]
Papers
- LeMat-GenBench: A Unified Evaluation Framework for Crystal Generative Models arXiv preprint, 2025 arXiv
- LeMat-Synth: A Multi-Modal Toolbox to Curate Broad Synthesis Procedure Databases from Scientific Literature NeurIPS 2025 Workshop on AI for Accelerated Materials Design arXiv
- AI for Scientific Discovery is a Social Problem arXiv preprint, 2025 arXiv
- Fact-Checking with Contextual Narratives: Leveraging Retrieval-Augmented LLMs for Social Media Analysis arXiv preprint, 2025 arXiv
- PSyDUCK: Training-Free Steganography for Latent Diffusion arXiv preprint, 2025 arXiv
-
LLM-Consensus: Multi-Agent Debate for Visual Misinformation Detection
ICML 2025 Workshop on Collaborative and Federated Agentic Workflows (Best Paper Award)
arXiv
- Toward Robust Real-World Audio Deepfake Detection: Closing the Explainability Gap ICML 2025 Workshop on Machine Learning for Audio
- Composable Workflow for Accelerating Neural Architecture Search Using In Situ Analytics for Protein Characterization International Conference on Parallel Processing (ICPP), 2023
- Spectral DefocusCam: Compressive Hyperspectral Imaging from Defocus Measurements AAAI Conference on Artificial Intelligence, 2022
Invited Talks
- MAD-Sherlock: Multi-Agent Debate for Visual Misinformation Detection Breaking Topics in AI Conference 2025, Imperial College London November 2025
- AI for Scientific Discovery is a Social Problem Toronto Data Workshop, Toronto Data Science Institute November 2025
- Cusp AI and the Future of Materials Discovery Fireside chat with Max Welling, World Summit AI 2025, Amsterdam October 2025
- Frontier Models and Human Values Panel with Val Bercovici, Marzieh Fadaee, and others, World Summit AI 2025, Amsterdam October 2025
- Superintelligence in Science Fireside chat with Ben Kompa (Lila Sciences), World Summit AI 2025, Amsterdam October 2025
- Deploying AI Outside the Lab: Authenticity, Safety, and Failure Modes Panel discussion, BBC Research & Development, London July 2025
- Strategies for Recruitment and Retention of Women in CS Grace Hopper Conference (GHC22) September 2022
Blogs
- SARLO-80: Worldwide Slant SAR Language Optic Dataset at 80 cm Resolution December 1, 2025
- The Pharmome Map: a comprehensive public dataset for drug-target interaction modeling November 18, 2025
- Advancing Predictive ADMET Modeling Through Community-Driven Science: The ExpansionRx-OpenADMET Blind Challenge October 27, 2025
- How to Train an Antibody Developability Model September 17, 2025
- SAIR: Accelerating Pharma R&D with AI-Powered Structural Intelligence September 2, 2025
- Bringing Fusion Down to Earth: ML for Stellarator Optimization July 2, 2025
- Accelerating AI for Drug Discovery: Ginkgo's GDPx Functional Genomics and GDPa Antibody Developability Dataset Series June 24, 2025
Awards
- Autonomous Intelligent Machines & Systems (AIMS) Scholar United Kingdom Research and Innovation Council
- Google DeepMind Scholarship Oxford Department of Computer Science
- Graduate Research Fellowship (NSF GRFP) National Science Foundation - declined
Community Activities
- Organizer Hugging Science - 7000+ strong community of OS contributors building in AI for Science
- Program Chair AI for Space Workshop at CVPR 2026
- Reviewer AAAI, NeurIPS Workshops