-
CryoDRGN-AI: Notes on ab initio heterogeneous reconstruction and its implications for spliceosome dynamics
A machine learning view of the CryoDRGN series, focusing on neural implicit representations, autodecoder-based ab initio reconstruction, physics-informed losses, and how latent structural landscapes can illuminate spliceosome dynamics.
-
Modeling life beyond standard Machine Learning algorithms
Why biological systems require multi-scale, strongly coupled, and perturbation-aware models, and how mathematical modeling and AI can move beyond standard CV and NLP paradigms.
-
Why the Diffusion Module matters in predictive models: a mathematical perspective
A mathematical analysis of why diffusion modules are important for modeling uncertainty, multimodal prediction, and high-dimensional output distributions.
-
Why C. elegans is a classic model for studying RNA Processing
A conceptual overview of SL trans-splicing, operon organization, and what the nematode C. elegans teaches us about RNA biology.