Quadratic models for understanding catapult dynamics of neural networks
- URL: http://arxiv.org/abs/2205.11787v3
- Date: Wed, 1 May 2024 23:15:37 GMT
- Title: Quadratic models for understanding catapult dynamics of neural networks
- Authors: Libin Zhu, Chaoyue Liu, Adityanarayanan Radhakrishnan, Mikhail Belkin,
- Abstract summary: We show that recently proposed Neural Quadratic Models can exhibit the "catapult phase" that arises when training such models with large learning rates.
Our analysis further demonstrates that quadratic models can be an effective tool for analysis of neural networks.
- Score: 15.381097076708535
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: While neural networks can be approximated by linear models as their width increases, certain properties of wide neural networks cannot be captured by linear models. In this work we show that recently proposed Neural Quadratic Models can exhibit the "catapult phase" [Lewkowycz et al. 2020] that arises when training such models with large learning rates. We then empirically show that the behaviour of neural quadratic models parallels that of neural networks in generalization, especially in the catapult phase regime. Our analysis further demonstrates that quadratic models can be an effective tool for analysis of neural networks.
Related papers
- Graph Neural Networks for Learning Equivariant Representations of Neural Networks [55.04145324152541]
We propose to represent neural networks as computational graphs of parameters.
Our approach enables a single model to encode neural computational graphs with diverse architectures.
We showcase the effectiveness of our method on a wide range of tasks, including classification and editing of implicit neural representations.
arXiv Detail & Related papers (2024-03-18T18:01:01Z) - A Survey on Statistical Theory of Deep Learning: Approximation, Training Dynamics, and Generative Models [13.283281356356161]
We review the literature on statistical theories of neural networks from three perspectives.
Results on excess risks for neural networks are reviewed.
Papers that attempt to answer how the neural network finds the solution that can generalize well on unseen data'' are reviewed.
arXiv Detail & Related papers (2024-01-14T02:30:19Z) - Six Lectures on Linearized Neural Networks [15.880450613409302]
We first recall the correspondence between neural networks and linear models via the so-called lazy regime.
We then review four models for linearized neural networks: linear regression with concentrated features, kernel ridge regression, random feature model and neural tangent model.
arXiv Detail & Related papers (2023-08-25T15:23:06Z) - How neural networks learn to classify chaotic time series [77.34726150561087]
We study the inner workings of neural networks trained to classify regular-versus-chaotic time series.
We find that the relation between input periodicity and activation periodicity is key for the performance of LKCNN models.
arXiv Detail & Related papers (2023-06-04T08:53:27Z) - Sparse deep neural networks for modeling aluminum electrolysis dynamics [0.5257115841810257]
We train sparse neural networks to model the system dynamics of an aluminum electrolysis simulator.
The sparse model structure has a significantly reduction in model complexity compared to a corresponding dense neural network.
The empirical study shows that the sparse models generalize better from small training sets than dense neural networks.
arXiv Detail & Related papers (2022-09-13T09:11:50Z) - EINNs: Epidemiologically-Informed Neural Networks [75.34199997857341]
We introduce a new class of physics-informed neural networks-EINN-crafted for epidemic forecasting.
We investigate how to leverage both the theoretical flexibility provided by mechanistic models as well as the data-driven expressability afforded by AI models.
arXiv Detail & Related papers (2022-02-21T18:59:03Z) - Sparse Flows: Pruning Continuous-depth Models [107.98191032466544]
We show that pruning improves generalization for neural ODEs in generative modeling.
We also show that pruning finds minimal and efficient neural ODE representations with up to 98% less parameters compared to the original network, without loss of accuracy.
arXiv Detail & Related papers (2021-06-24T01:40:17Z) - The Neural Coding Framework for Learning Generative Models [91.0357317238509]
We propose a novel neural generative model inspired by the theory of predictive processing in the brain.
In a similar way, artificial neurons in our generative model predict what neighboring neurons will do, and adjust their parameters based on how well the predictions matched reality.
arXiv Detail & Related papers (2020-12-07T01:20:38Z) - A Bayesian Perspective on Training Speed and Model Selection [51.15664724311443]
We show that a measure of a model's training speed can be used to estimate its marginal likelihood.
We verify our results in model selection tasks for linear models and for the infinite-width limit of deep neural networks.
Our results suggest a promising new direction towards explaining why neural networks trained with gradient descent are biased towards functions that generalize well.
arXiv Detail & Related papers (2020-10-27T17:56:14Z) - Learning Queuing Networks by Recurrent Neural Networks [0.0]
We propose a machine-learning approach to derive performance models from data.
We exploit a deterministic approximation of their average dynamics in terms of a compact system of ordinary differential equations.
This allows for an interpretable structure of the neural network, which can be trained from system measurements to yield a white-box parameterized model.
arXiv Detail & Related papers (2020-02-25T10:56:47Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.