Summary of Strategic Usage in a Multi-learner Setting, by Eliot Shekhtman and Sarah Dean
Strategic Usage in a Multi-Learner Setting
by Eliot Shekhtman, Sarah Dean
First submitted to arxiv on: 29 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Science and Game Theory (cs.GT)
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 explores the strategic behavior of users in online learning systems where multiple services compete for user attention. In this setting, users may choose which services to use based on their individual goals, influencing the data that services collect and utilize. The authors analyze a scenario where users select among several available services to achieve desired classifications while services strive to minimize losses. They demonstrate that naive retraining can lead to oscillation even with all users observed at different times, but using memory of past observations ensures convergent behavior for certain loss function classes. The results are validated through synthetic and real-world data. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine online learning platforms where people choose between different services based on what they want to achieve. Services use this user data to improve their performance. In reality, users might strategically select which services to use to get the outcome they desire. This can affect how well the services perform and whether they collect accurate information about users. The paper examines a situation where users choose among multiple services to get the right classification while services try to minimize losses. It shows that simply retraining the model without considering past user behavior can lead to fluctuations, but using previous observations helps ensure the model improves over time. |
Keywords
* Artificial intelligence * Attention * Classification * Loss function * Online learning
