PEP: Parameter Ensembling by Perturbation

• URL: http://arxiv.org/abs/2010.12721v1
• Date: Sat, 24 Oct 2020 00:16:03 GMT
• Title: PEP: Parameter Ensembling by Perturbation
• Authors: Alireza Mehrtash, Purang Abolmaesumi, Polina Golland, Tina Kapur, Demian Wassermann, William M. Wells III
• Abstract summary: Ensembling by Perturbation (PEP) constructs an ensemble of parameter values as random perturbations of the optimal parameter set from training. PEP provides a small improvement in performance, and, in some cases, a substantial improvement in empirical calibration. PEP can be used to probe the level of overfitting that occurred during training.
• Score: 13.221295194854642
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Ensembling is now recognized as an effective approach for increasing the predictive performance and calibration of deep networks. We introduce a new approach, Parameter Ensembling by Perturbation (PEP), that constructs an ensemble of parameter values as random perturbations of the optimal parameter set from training by a Gaussian with a single variance parameter. The variance is chosen to maximize the log-likelihood of the ensemble average ($\mathbb{L}$) on the validation data set. Empirically, and perhaps surprisingly, $\mathbb{L}$ has a well-defined maximum as the variance grows from zero (which corresponds to the baseline model). Conveniently, calibration level of predictions also tends to grow favorably until the peak of $\mathbb{L}$ is reached. In most experiments, PEP provides a small improvement in performance, and, in some cases, a substantial improvement in empirical calibration. We show that this "PEP effect" (the gain in log-likelihood) is related to the mean curvature of the likelihood function and the empirical Fisher information. Experiments on ImageNet pre-trained networks including ResNet, DenseNet, and Inception showed improved calibration and likelihood. We further observed a mild improvement in classification accuracy on these networks. Experiments on classification benchmarks such as MNIST and CIFAR-10 showed improved calibration and likelihood, as well as the relationship between the PEP effect and overfitting; this demonstrates that PEP can be used to probe the level of overfitting that occurred during training. In general, no special training procedure or network architecture is needed, and in the case of pre-trained networks, no additional training is needed.

Related papers

• Just How Flexible are Neural Networks in Practice? [89.80474583606242]
  It is widely believed that a neural network can fit a training set containing at least as many samples as it has parameters. In practice, however, we only find solutions via our training procedure, including the gradient and regularizers, limiting flexibility.
  arXiv  Detail & Related papers  (2024-06-17T12:24:45Z)
• Domain-adaptive and Subgroup-specific Cascaded Temperature Regression for Out-of-distribution Calibration [16.930766717110053]
  We propose a novel meta-set-based cascaded temperature regression method for post-hoc calibration. We partition each meta-set into subgroups based on predicted category and confidence level, capturing diverse uncertainties. A regression network is then trained to derive category-specific and confidence-level-specific scaling, achieving calibration across meta-sets.
  arXiv  Detail & Related papers  (2024-02-14T14:35:57Z)
• Sparse is Enough in Fine-tuning Pre-trained Large Language Models [98.46493578509039]
  We propose a gradient-based sparse fine-tuning algorithm, named Sparse Increment Fine-Tuning (SIFT) We validate its effectiveness on a range of tasks including the GLUE Benchmark and Instruction-tuning.
  arXiv  Detail & Related papers  (2023-12-19T06:06:30Z)
• Variational Linearized Laplace Approximation for Bayesian Deep Learning [11.22428369342346]
  We propose a new method for approximating Linearized Laplace Approximation (LLA) using a variational sparse Gaussian Process (GP) Our method is based on the dual RKHS formulation of GPs and retains, as the predictive mean, the output of the original DNN. It allows for efficient optimization, which results in sub-linear training time in the size of the training dataset.
  arXiv  Detail & Related papers  (2023-02-24T10:32:30Z)
• Reliable Prediction Intervals with Directly Optimized Inductive Conformal Regression for Deep Learning [3.42658286826597]
  Predictions intervals (PIs) are used to quantify the uncertainty of each prediction in deep learning regression. Many approaches to improve the quality of PIs can effectively reduce the width of PIs, but they do not ensure that enough real labels are captured. In this study, we use Directly Optimized Inductive Conformal Regression (DOICR) that takes only the average width of PIs as the loss function. Benchmark experiments show that DOICR outperforms current state-of-the-art algorithms for regression problems.
  arXiv  Detail & Related papers  (2023-02-02T04:46:14Z)
• Predicting Deep Neural Network Generalization with Perturbation Response Curves [58.8755389068888]
  We propose a new framework for evaluating the generalization capabilities of trained networks. Specifically, we introduce two new measures for accurately predicting generalization gaps. We attain better predictive scores than the current state-of-the-art measures on a majority of tasks in the Predicting Generalization in Deep Learning (PGDL) NeurIPS 2020 competition.
  arXiv  Detail & Related papers  (2021-06-09T01:37:36Z)
• Efficient training of physics-informed neural networks via importance sampling [2.9005223064604078]
  Physics-In Neural Networks (PINNs) are a class of deep neural networks that are trained to compute systems governed by partial differential equations (PDEs) We show that an importance sampling approach will improve the convergence behavior of PINNs training.
  arXiv  Detail & Related papers  (2021-04-26T02:45:10Z)
• Exploring the Uncertainty Properties of Neural Networks' Implicit Priors in the Infinite-Width Limit [47.324627920761685]
  We use recent theoretical advances that characterize the function-space prior to an ensemble of infinitely-wide NNs as a Gaussian process. This gives us a better understanding of the implicit prior NNs place on function space. We also examine the calibration of previous approaches to classification with the NNGP.
  arXiv  Detail & Related papers  (2020-10-14T18:41:54Z)
• Evaluating Prediction-Time Batch Normalization for Robustness under Covariate Shift [81.74795324629712]
  We call prediction-time batch normalization, which significantly improves model accuracy and calibration under covariate shift. We show that prediction-time batch normalization provides complementary benefits to existing state-of-the-art approaches for improving robustness. The method has mixed results when used alongside pre-training, and does not seem to perform as well under more natural types of dataset shift.
  arXiv  Detail & Related papers  (2020-06-19T05:08:43Z)
• Extrapolation for Large-batch Training in Deep Learning [72.61259487233214]
  We show that a host of variations can be covered in a unified framework that we propose. We prove the convergence of this novel scheme and rigorously evaluate its empirical performance on ResNet, LSTM, and Transformer.
  arXiv  Detail & Related papers  (2020-06-10T08:22:41Z)
• Scaling Equilibrium Propagation to Deep ConvNets by Drastically Reducing its Gradient Estimator Bias [65.13042449121411]
  In practice, training a network with the gradient estimates provided by EP does not scale to visual tasks harder than MNIST. We show that a bias in the gradient estimate of EP, inherent in the use of finite nudging, is responsible for this phenomenon. We apply these techniques to train an architecture with asymmetric forward and backward connections, yielding a 13.2% test error.
  arXiv  Detail & Related papers  (2020-06-06T09:36:07Z)

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.