Summary of Unsupervised Domain Adaptation Via Data Pruning, by Andrea Napoli et al.
Unsupervised Domain Adaptation Via Data Pruning
by Andrea Napoli, Paul White
First submitted to arxiv on: 18 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- 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 removal of carefully-selected examples from training data has been found to improve the robustness of machine learning models. However, selecting the best examples remains an open question. This paper proposes AdaPrune, a method for unsupervised domain adaptation (UDA) that removes training examples to align the training distribution with the target data. The problem is formulated and solved as an integer quadratic program using maximum mean discrepancy (MMD) as the alignment criterion. Experimental results on a bioacoustic event detection task show that AdaPrune outperforms related techniques, including CORAL, and provides a principled approach to data pruning. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine learning models can get better by removing some examples from their training data. But which ones should we remove? This paper talks about unsupervised domain adaptation (UDA), which helps machines learn to work with new types of data they haven’t seen before. The authors propose a way to do this called AdaPrune, which removes examples that are different from the target data. They use a special math formula called maximum mean discrepancy (MMD) to figure out which examples to remove. By doing so, AdaPrune can help machines make more accurate predictions. |
Keywords
» Artificial intelligence » Alignment » Domain adaptation » Event detection » Machine learning » Pruning » Unsupervised
