Summary of Clce: An Approach to Refining Cross-entropy and Contrastive Learning For Optimized Learning Fusion, by Zijun Long and George Killick and Lipeng Zhuang and Gerardo Aragon-camarasa and Zaiqiao Meng and Richard Mccreadie
CLCE: An Approach to Refining Cross-Entropy and Contrastive Learning for Optimized Learning Fusion
by Zijun Long, George Killick, Lipeng Zhuang, Gerardo Aragon-Camarasa, Zaiqiao Meng, Richard Mccreadie
First submitted to arxiv on: 22 Feb 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper introduces a novel approach to pre-trained image models called CLCE, which combines Label-Aware Contrastive Learning with Cross-Entropy loss. This approach aims to improve model generalization and stability by leveraging hard negative mining and reducing the dependency on large batch sizes. The authors demonstrate that CLCE outperforms traditional CE in Top-1 accuracy across twelve benchmarks, achieving gains of up to 3.52% in few-shot learning scenarios and 3.41% in transfer learning settings using the BEiT-3 model. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates a new way to train image models called CLCE. It combines two methods: Label-Aware Contrastive Learning and Cross-Entropy loss. This helps the model learn better and be more stable. The results show that CLCE is better than usual methods in many cases, especially when learning from small amounts of data or transferring knowledge between tasks. |
Keywords
* Artificial intelligence * Cross entropy * Few shot * Generalization * Transfer learning
