Which biological relationships can be compressed without losing the information needed for scientific inference?
The research thesis
My proposed direction extends the probabilistic modeling of TranshumanGene and AI4Omics into quantum-inspired representations. Tensor networks and structured optimization offer a language for exploring correlations, constrained selection and compact models of biological networks.
Quantum-inspired methods can run on classical computers. Their connection to quantum mathematics does not imply execution on a quantum processor. This distinction makes the computational experiment clear and allows useful work with ordinary scientific infrastructure.
A first experiment
Start with a synthetic gene-regulatory or metabolic network whose important relationships are known. Compare a compressed representation with a full small model and a conventional low-rank baseline.
Measure reconstruction error, memory, runtime and stability when interactions or observations are missing. A second experiment could test constrained feature selection on a public, properly scoped research dataset.
What a partner would receive
A reproducible modeling study describing which correlations survive compression, which disappear and how those changes affect inference. The intended outputs are research models and computational evidence.
Scientific connection
The bridge is between my work in genomics, probabilistic inference and complex systems, and mathematical techniques developed for quantum many-body problems. It is an exploration of representation and computation, with any practical performance benefit left to the experiment.
Sources & scientific context
Author
Maurizio VivianiIndependent research · Robotics
Take one question into an experiment.
Discuss a focused research collaboration, benchmark or implementation study.
Discuss this direction