Summary of A Comprehensive Sustainable Framework For Machine Learning and Artificial Intelligence, by Roberto Pagliari et al.
A Comprehensive Sustainable Framework for Machine Learning and Artificial Intelligence
by Roberto Pagliari, Peter Hill, Po-Yu Chen, Maciej Dabrowny, Tingsheng Tan, Francois Buet-Golfouse
First submitted to arxiv on: 17 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: 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 In this paper, researchers introduce a new framework for Sustainable Machine Learning that incorporates four key pillars: fairness, privacy, interpretability, and greenhouse gas emissions. While past literature has addressed these pillars individually, this work is the first to consider all of them simultaneously. The authors propose FPIG, a general AI pipeline that allows users to learn trade-offs between the pillars. A meta-learning algorithm is developed to estimate the four key pillars given dataset summary, model architecture, and hyperparameters before training. The algorithm enables users to select the optimal model architecture for a given dataset and set of user requirements on the pillars. This work demonstrates the trade-offs under the FPIG framework on three classical datasets and showcases how it can aid model selection. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps make artificial intelligence more sustainable by looking at four important areas: fairness, privacy, understanding what’s happening in the model, and reducing carbon emissions. Past research has focused on one or two of these areas separately, but this study is the first to consider all four together. The authors propose a new approach that lets users see how different choices affect the results and make better decisions about which models to use. This could be helpful for people who want to use AI in real-world applications. |
Keywords
» Artificial intelligence » Machine learning » Meta learning
