Summary of Domain-agnostic Mutual Prompting For Unsupervised Domain Adaptation, by Zhekai Du et al.
Domain-Agnostic Mutual Prompting for Unsupervised Domain Adaptation
by Zhekai Du, Xinyao Li, Fengling Li, Ke Lu, Lei Zhu, Jingjing Li
First submitted to arxiv on: 5 Mar 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 proposed Domain-Agnostic Mutual Prompting (DAMP) technique aims to bridge the gap in conventional Unsupervised Domain Adaptation (UDA) by leveraging pre-trained vision-language models. Current methods struggle to handle complex domain shifts and learn textual prompts separately for source and target domains, limiting cross-domain knowledge transfer. DAMP mutually aligns visual and textual embeddings by utilizing image contextual information to prompt the language branch in a domain-agnostic way. The approach is learned through a cross-attention module and regularized with semantic-consistency and instance-discrimination losses. Experimental results on three UDA benchmarks demonstrate the superiority of DAMP over state-of-the-art approaches. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary DAMP is a new method that helps computers learn from different types of data without needing lots of extra training. Right now, computers have trouble learning from different kinds of data, like pictures or words. The problem is that they don’t understand what’s important and what’s not. DAMP fixes this by teaching the computer to look at both the picture and the text together. This helps the computer learn more about the things it doesn’t know yet. |
Keywords
» Artificial intelligence » Cross attention » Domain adaptation » Prompt » Prompting » Unsupervised
