Summary of Stochastic Online Conformal Prediction with Semi-bandit Feedback, by Haosen Ge et al.
Stochastic Online Conformal Prediction with Semi-Bandit Feedback
by Haosen Ge, Hamsa Bastani, Osbert Bastani
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 proposed conformal prediction algorithm modifies a model to output sets of labels instead of single labels, providing uncertainty quantification. The algorithm is designed for the online learning setting where examples arrive over time, with semi-bandit feedback allowing only true label observation if it’s contained in the prediction set. This is beneficial in applications like document retrieval calibration or image classification. Compared to optimal conformal predictors, the proposed algorithm achieves sublinear regret. It is evaluated on retrieval, image classification, and auction price-setting tasks, demonstrating good performance compared to several baselines. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In this paper, scientists are working on a new way to make computer models more reliable by giving them “uncertainty” – basically, a range of possible answers instead of just one. This is useful when we’re trying to do things like find the best documents or classify images. The researchers have come up with a special algorithm that can do this online, meaning it can learn and adapt as new data comes in. They tested their idea on several tasks and found it worked well. |
Keywords
» Artificial intelligence » Image classification » Online learning
