p-adic Bi-Filtrations for Topological Machine Learning on Genomic Sequences

pVR is an alignment-free genomic classification framework that combines p-adic hierarchical structure and k-mer composition within a bi-filtered Vietoris–Rips complex. It provides stability and prime-invariance guarantees, extracts topological features for machine learning, and achieves strong performance on several low-sample genomic benchmarks, outperforming established baselines and frozen foundation-model embeddings.

Recommended citation: Dash, T., & Sachdeva, G. (2026). p-adic Bi-Filtrations for Topological Machine Learning on Genomic Sequences. arXiv preprint arXiv:2606.06117.
Download Paper