A multi-level collaborative self-distillation learning for improving adaptive inference efficiency

A multi-level collaborative self-distillation learning for improving adaptive inference efficiency

Abstract A multi-exit network is an important technique for achieving adaptive inference by dynamically allocating computational resources based on different input samples. The existing works mainly treat the final classifier as the teacher, enhancing the classification accuracy by transferring know...

Full description

Bibliographic Details
Main Authors: Likun Zhang, Jinbao Li, Benqian Zhang, Yahong Guo
Format: Article
Language:English
Published: Springer 2024-08-01
Series:Complex & Intelligent Systems
Subjects:
Multi-level collaborative self-distillation
Multi-exit network
Collaborative learning
Adaptive inference
Efficient computing
Online Access:https://doi.org/10.1007/s40747-024-01572-3

Internet

https://doi.org/10.1007/s40747-024-01572-3

Similar Items