A Meta-Analysis of Overfitting in Machine Learning

By Rebecca Roelofs, Vaishaal Shankar, B. Recht, Sara Fridovich-Keil, Moritz Hardt, John Miller, Ludwig Schmidt. 2019

198 citations recorded by the source corpus.

Subfields: Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Evaluation & Benchmarks

1. John Miller
2. Ludwig Schmidt
3. Rohan Taori
4. B. Recht
5. Jack Hessel
6. Karl Krauth
7. Herv'e J'egou
8. Hugo Touvron

Citations in the corpus

• The Effect of Natural Distribution Shift on Question Answering Models (2020, background): “Furthermore, Roelofs et al. (2019) also found little to no evidence of adaptive overfitting in a host of classification competitions on the Kaggle platform.”
• VisIT-Bench: A Benchmark for Vision-Language Instruction Following Inspired by Real-World Use (2023, background): “Recent work has indicated that improvements on these high-quality benchmark datasets is not the result of overfitting, and is a reliable indicator of genuine progress beyond the benchmark data [77, 78, 79, 80].”