Summary of Microfoundation Inference For Strategic Prediction, by Daniele Bracale et al.
Microfoundation Inference for Strategic Prediction
by Daniele Bracale, Subha Maity, Felipe Maia Polo, Seamus Somerstep, Moulinath Banerjee, Yuekai Sun
First submitted to arxiv on: 13 Nov 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Machine Learning (cs.LG); Methodology (stat.ME)
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 methodology aims to address the issue of performative prediction in machine learning, where predictive models influence the target variable’s distribution. This phenomenon is often driven by strategic actions from stakeholders with vested interests. To tackle this challenge, the authors suggest an approach for learning the long-term impacts of predictive models on the population. The method models agents’ responses as a cost-adjusted utility maximization problem and proposes estimates for these costs. Optimal transport is leveraged to align pre-model exposure (ex ante) and post-model exposure (ex post) distributions. The authors provide a rate of convergence for this proposed estimate and demonstrate its quality through empirical experiments on a credit-scoring dataset. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores how predictive models can shape the world by influencing people’s decisions. When we use machines to make predictions, it can affect what people do in the future. This is called performative prediction. The problem is that most researchers don’t realize this is happening and don’t think about the long-term effects. To fix this, the authors suggest a new way to understand how predictive models will change the world. They use math from economics to model how people make decisions based on predictions. Then, they use a technique called optimal transport to connect the past and future in a special way. The authors show that their approach works well using real data from credit scoring. |
Keywords
» Artificial intelligence » Machine learning
