arXivAI Research
Designing Unstructured Proteins: IDiom and RL-SAE Enable Interpretable Feature Control
突破無結構蛋白設計瓶頸！IDiom 與 RL-SAE 開創可解釋性的蛋白質功能定製
Researchers developed IDiom, a protein language model for intrinsically disordered regions, alongside RL-SAE, a reinforcement learning method that uses sparse autoencoder features to precisely control engineered functions.
2 min read