Protein Structure Prediction in the 3D HP Model Using Deep Reinforcement Learning
This work addresses protein folding for computational biology, presenting incremental improvements in training efficiency.
The paper tackles protein structure prediction in the 3D HP model by introducing two deep learning architectures, achieving optimal conformations with 25% fewer training episodes for short sequences and matching best-known energy values for longer ones.
We address protein structure prediction in the 3D Hydrophobic-Polar lattice model through two novel deep learning architectures. For proteins under 36 residues, our hybrid reservoir-based model combines fixed random projections with trainable deep layers, achieving optimal conformations with 25% fewer training episodes. For longer sequences, we employ a long short-term memory network with multi-headed attention, matching best-known energy values. Both architectures leverage a stabilized Deep Q-Learning framework with experience replay and target networks, demonstrating consistent achievement of optimal conformations while significantly improving training efficiency compared to existing methods.