Policy-transferred reinforcement learning enables sample-efficient inverse design of thermally active composites with programmable 4D shape transformation.
Share your take
Add a clinician perspective alongside expert commentary.
Lee et al. (2026) studied this question.