Kevin K. Yang: Engineering Proteins with ML
The Gradient: Perspectives on AI
00:00
Exploring Pre-Trained Protein Language Model Transfer Learning
They discuss the effectiveness of middle layer outputs from pre-trained models and how they can be as effective as the final layer representations. They also explore how a middle layer is often better for engineering tasks and different training pre-training checkpoints can be equally or even more optimal.
Transcript
Play full episode
