Shifts protein fitness optimization from continuous embeddings to discrete Quadratic Unconstrained Binary Optimization (QUBO).
March 31, 2026
Original Paper
Q-BIOLAT: Binary Latent Protein Fitness Landscapes for QUBO-Based Optimization
arXiv · 2603.27526
The Takeaway
It challenges the reliance on continuous latent spaces for protein modeling, showing that structured binary spaces allow classical combinatorial and quantum-inspired optimization to outperform standard methods. It provides a blueprint for connecting ML protein models with more efficient discrete search algorithms.
From the abstract
Protein fitness optimization is inherently a discrete combinatorial problem, yet most learning-based approaches rely on continuous representations and are primarily evaluated through predictive accuracy. We introduce Q-BIOLAT, a framework for modeling and optimizing protein fitness landscapes in compact binary latent spaces. Starting from pretrained protein language model embeddings, we construct binary latent representations and learn a quadratic unconstrained binary optimization (QUBO) surroga