Summary of Becotta: Input-dependent Online Blending Of Experts For Continual Test-time Adaptation, by Daeun Lee et al.
BECoTTA: Input-dependent Online Blending of Experts for Continual Test-time Adaptation
by Daeun Lee, Jaehong Yoon, Sung Ju Hwang
First submitted to arxiv on: 13 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 This paper proposes BECoTTA, a modular framework for Continual Test Time Adaptation (CTTA) that efficiently adapts to continuous unseen domains while retaining previously learned knowledge. The proposed framework, Mixture-of Domain Low-rank Experts (MoDE), consists of two core components: Domain-Adaptive Routing and Domain-Expert Synergy Loss. These components enable the selective capture of domain-adaptive knowledge and maximize the dependency between each domain and expert. The authors validate that BECoTTA outperforms multiple CTTA scenarios, including disjoint and gradual domain shifts, while requiring only ~98% fewer trainable parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary BECoTTA is a new way to help machines learn new things without forgetting what they already know. Right now, it’s hard to teach machines to adapt to changes in the world, but this method makes it easier. It works by using special “experts” that can learn from different areas of expertise and combining them to make better decisions. The results show that BECoTTA is more effective than other methods at adapting to changing situations while still remembering what it already knows. |
