Summary of Misconduct in Post-selections and Deep Learning, by Juyang Weng
Misconduct in Post-Selections and Deep Learning
by Juyang Weng
First submitted to arxiv on: 13 Feb 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 This theoretical paper investigates misconduct in Deep Learning and post-selection methods. The authors identify cheating and hiding as common practices, where models are trained without a test and bad-looking data is concealed. The study highlights the importance of reporting average errors on validation sets and ranked error positions to prevent these misdeeds. The research reveals that cross-validation on data splits does not necessarily exonerate post-selection methods from misconduct, indicating that these statistical learners based on validation set errors are statistically invalid. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how machine learning models can be unfair and misleading. It shows that many models cheat by hiding their mistakes or using fake data to make them look better. The researchers think that we need more transparency in how models are tested, so people know what’s really going on. They also found that even when models are tested on different parts of the same data, they can still be misleading and unreliable. |
Keywords
* Artificial intelligence * Deep learning * Machine learning
