-
Lower Bounds on the Generalization Error of Nonlinear Learning Models
We study in this paper lower bounds for the generalization error of mode...
read it
-
Dimension Independent Generalization Error with Regularized Online Optimization
One classical canon of statistics is that large models are prone to over...
read it
-
The Neural Tangent Kernel in High Dimensions: Triple Descent and a Multi-Scale Theory of Generalization
Modern deep learning models employ considerably more parameters than req...
read it
-
The Slow Deterioration of the Generalization Error of the Random Feature Model
The random feature model exhibits a kind of resonance behavior when the ...
read it
-
An Investigation of Why Overparameterization Exacerbates Spurious Correlations
We study why overparameterization – increasing model size well beyond th...
read it
-
The generalization error of max-margin linear classifiers: High-dimensional asymptotics in the overparametrized regime
Modern machine learning models are often so complex that they achieve va...
read it
-
Big Neural Networks Waste Capacity
This article exposes the failure of some big neural networks to leverage...
read it
High-dimensional dynamics of generalization error in neural networks
We perform an average case analysis of the generalization dynamics of large neural networks trained using gradient descent. We study the practically-relevant "high-dimensional" regime where the number of free parameters in the network is on the order of or even larger than the number of examples in the dataset. Using random matrix theory and exact solutions in linear models, we derive the generalization error and training error dynamics of learning and analyze how they depend on the dimensionality of data and signal to noise ratio of the learning problem. We find that the dynamics of gradient descent learning naturally protect against overtraining and overfitting in large networks. Overtraining is worst at intermediate network sizes, when the effective number of free parameters equals the number of samples, and thus can be reduced by making a network smaller or larger. Additionally, in the high-dimensional regime, low generalization error requires starting with small initial weights. We then turn to non-linear neural networks, and show that making networks very large does not harm their generalization performance. On the contrary, it can in fact reduce overtraining, even without early stopping or regularization of any sort. We identify two novel phenomena underlying this behavior in overcomplete models: first, there is a frozen subspace of the weights in which no learning occurs under gradient descent; and second, the statistical properties of the high-dimensional regime yield better-conditioned input correlations which protect against overtraining. We demonstrate that naive application of worst-case theories such as Rademacher complexity are inaccurate in predicting the generalization performance of deep neural networks, and derive an alternative bound which incorporates the frozen subspace and conditioning effects and qualitatively matches the behavior observed in simulation.
READ FULL TEXT
Comments
There are no comments yet.