Summary of Generalization Bounds For Dependent Data Using Online-to-batch Conversion, by Sagnik Chatterjee et al.
Generalization Bounds for Dependent Data using Online-to-Batch Conversion
by Sagnik Chatterjee, Manuj Mukherjee, Alhad Sethi
First submitted to arxiv on: 22 May 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 paper presents a new approach to bounding the generalization error of batch learning algorithms trained on dependent data. Unlike previous works, it does not require any stability assumptions and can be applied to any batch learning algorithm. The authors use the Online-to-Batch (OTB) conversion framework to shift the burden of stability from the batch learner to an artificially constructed online learner. They introduce a new notion of algorithmic stability based on Wasserstein distances and show that the EWA algorithm satisfies this stability condition. The authors then instantiate their bounds using the EWA algorithm. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how well machine learning models work when trained on data that is connected in some way, like a sequence or a graph. They want to know how well these models generalize to new, unseen data. To do this, they use a framework called Online-to-Batch (OTB) that lets them turn an online learning algorithm into a batch one. This helps them get around the need for special assumptions about the stability of the model. The authors also come up with a new way to measure how stable an online learning algorithm is and show that a popular algorithm called EWA meets this standard. |
Keywords
» Artificial intelligence » Generalization » Machine learning » Online learning
