Abstract | ||
---|---|---|
We empirically show that the test error of deep networks can be estimated by simply training the same architecture on the same training set but with a different run of Stochastic Gradient Descent (SGD), and measuring the disagreement rate between the two networks on unlabeled test data. This builds on -- and is a stronger version of -- the observation in Nakkiran & Bansal '20, which requires the second run to be on an altogether fresh training set. We further theoretically show that this peculiar phenomenon arises from the \emph{well-calibrated} nature of \emph{ensembles} of SGD-trained models. This finding not only provides a simple empirical measure to directly predict the test error using unlabeled test data, but also establishes a new conceptual connection between generalization and calibration. |
Year | Venue | Keywords |
---|---|---|
2022 | International Conference on Learning Representations (ICLR) | Generalization,Deep Learning,Empirical Phenomenon,Accuracy Estimation,Stochastic Gradient Descent |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yiding Jiang | 1 | 3 | 2.41 |
Vaishnavh Nagarajan | 2 | 0 | 1.35 |
Christina Baek | 3 | 0 | 0.34 |
J. Zico Kolter | 4 | 1270 | 84.23 |