Summary of Dataset Condensation with Latent Quantile Matching, by Wei Wei et al.
Dataset Condensation with Latent Quantile Matching
by Wei Wei, Tom De Schepper, Kevin Mets
First submitted to arxiv on: 14 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); 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 a new method called Latent Quantile Matching (LQM) to improve dataset condensation (DC) methods, which aim to accelerate machine learning model training by learning smaller synthesized datasets. Current DC methods rely on distribution matching (DM), but this approach has limitations, including weak matching power and lack of outlier regularization. LQM addresses these shortcomings by matching the quantiles of latent embeddings to minimize the goodness of fit test statistic between two distributions. The proposed method outperforms previous state-of-the-art DM-based DC approaches on both image and graph-structured datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to make machine learning models train faster. Right now, scientists are trying to create smaller versions of big datasets that can help train these models. They’re using something called distribution matching to do this, but it has some problems. The new method they propose, called Latent Quantile Matching, fixes these issues by looking at how the data is distributed in a special way. This helps make better small datasets that can train models more efficiently. |
Keywords
* Artificial intelligence * Machine learning * Regularization
