- All the latest frameworks, architectures and ideas for data processing and inference.
- Leveraging generative/inferential techniques to create interpretable symbolic systems.
I'm interested in:
- Data, decisions and operations integration software for medical institutions.
- Personal health systems for everyone.
- AI assistants made for medical professionals, hospitals and practices.
- Embodied and multimodal AI with internal world models.
Making:
- Evaluating and detecting statistical invalidity in the current medical literature and that can generate new insights.
- Ensuring that algorithms can explain their decisions (or we understand their decision-making; mechanistic interpretability).
- Ensuring patient data is secure and that systems are resistant to adversarial attacks by misaligned agents.
Systems: