Summary of Aggregate Representation Measure For Predictive Model Reusability, by Vishwesh Sangarya and Richard Bradford and Jung-eun Kim
Aggregate Representation Measure for Predictive Model Reusability
by Vishwesh Sangarya, Richard Bradford, Jung-Eun Kim
First submitted to arxiv on: 15 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Computers and Society (cs.CY)
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 Aggregated Representation Measure (ARM) is a predictive quantifier that estimates the retraining cost of a trained model in distribution shifts. ARM quantifies the change in the model’s representation from old to new data, providing a single concise index of resources required for retraining. This enables reuse of models with lower costs than training new ones from scratch. The results show ARM reasonably predicts retraining costs for varying noise intensities and allows comparisons among multiple architectures to determine the most cost-effective option. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper proposes a new way to predict how much it will cost to update a trained model when the data changes. They call this prediction “retraining cost” and they want to know how much energy, time, and carbon emissions are needed to do this update. The method they use is called Aggregated Representation Measure (ARM). ARM looks at how the model’s understanding of the data changes from old to new data, and it gives a single number that shows how much resources will be needed for retraining. |
