expert pruning
Description: A model-compression technique that removes or disables selected expert components from a mixture-of-experts neural network while aiming to preserve performance.
Primary topic: Advanced Neural Network Applications (60%)
OpenAlex ID: keywords/expert-pruning
