/thumbnails/rl-dist.jpg

Distributional Reinforcement Learning in High-Dimensional Spaces

An analysis of distributional value functions when applied to continuous control tasks with sparse rewards.

/thumbnails/tda-nn.jpg

Topological Data Analysis for Neural Network Interpretability

Using persistent homology to understand the shape of activation manifolds in deep classification networks.

/thumbnails/genomic-transformer.jpg

Efficient Transformers for Long-Context Genomic Sequencing

Modifying attention mechanisms to handle DNA sequences of up to 100k base pairs with linear complexity.