Robust Training of Neural Networks at Arbitrary Precision and Sparsity

• URL: http://arxiv.org/abs/2409.09245v1
• Date: Sat, 14 Sep 2024 00:57:32 GMT
• Title: Robust Training of Neural Networks at Arbitrary Precision and Sparsity
• Authors: Chengxi Ye, Grace Chu, Yanfeng Liu, Yichi Zhang, Lukasz Lew, Andrew Howard,
• Abstract summary: The discontinuous operations inherent in quantization and sparsification introduce obstacles to backpropagation. This is particularly challenging when training deep neural networks in ultra-low precision and sparse regimes. We propose a novel, robust, and universal solution: a denoising affine transform.
• Score: 11.177990498697845
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The discontinuous operations inherent in quantization and sparsification introduce obstacles to backpropagation. This is particularly challenging when training deep neural networks in ultra-low precision and sparse regimes. We propose a novel, robust, and universal solution: a denoising affine transform that stabilizes training under these challenging conditions. By formulating quantization and sparsification as perturbations during training, we derive a perturbation-resilient approach based on ridge regression. Our solution employs a piecewise constant backbone model to ensure a performance lower bound and features an inherent noise reduction mechanism to mitigate perturbation-induced corruption. This formulation allows existing models to be trained at arbitrarily low precision and sparsity levels with off-the-shelf recipes. Furthermore, our method provides a novel perspective on training temporal binary neural networks, contributing to ongoing efforts to narrow the gap between artificial and biological neural networks.

Related papers

• Progressive Element-wise Gradient Estimation for Neural Network Quantization [2.1413624861650358]
  Quantization-Aware Training (QAT) methods rely on the Straight-Through Estimator (STE) to address the non-differentiability of discretization functions.<n>We propose Progressive Element-wise Gradient Estimation (PEGE) to address discretization errors between continuous and quantized values.<n>PEGE consistently outperforms existing backpropagation methods and enables low-precision models to match or even outperform the accuracy of their full-precision counterparts.
  arXiv  Detail & Related papers  (2025-08-27T15:59:36Z)
• Certified Neural Approximations of Nonlinear Dynamics [52.79163248326912]
  In safety-critical contexts, the use of neural approximations requires formal bounds on their closeness to the underlying system.<n>We propose a novel, adaptive, and parallelizable verification method based on certified first-order models.
  arXiv  Detail & Related papers  (2025-05-21T13:22:20Z)
• Adaptive Class Emergence Training: Enhancing Neural Network Stability and Generalization through Progressive Target Evolution [0.0]
  We propose a novel training methodology for neural networks in classification problems. We evolve the target outputs from a null vector to one-hot encoded vectors throughout the training process. This gradual transition allows the network to adapt more smoothly to the increasing complexity of the classification task.
  arXiv  Detail & Related papers  (2024-09-04T03:25:48Z)
• QGait: Toward Accurate Quantization for Gait Recognition with Binarized Input [17.017127559393398]
  We propose a differentiable soft quantizer, which better simulates the gradient of the round function during backpropagation. This enables the network to learn from subtle input perturbations. We further refine the training strategy to ensure convergence while simulating quantization errors.
  arXiv  Detail & Related papers  (2024-05-22T17:34:18Z)
• Robust Stochastically-Descending Unrolled Networks [85.6993263983062]
  Deep unrolling is an emerging learning-to-optimize method that unrolls a truncated iterative algorithm in the layers of a trainable neural network. We show that convergence guarantees and generalizability of the unrolled networks are still open theoretical problems. We numerically assess unrolled architectures trained under the proposed constraints in two different applications.
  arXiv  Detail & Related papers  (2023-12-25T18:51:23Z)
• Tight and Efficient Gradient Bounds for Parameterized Quantum Circuits [7.0379869298557844]
  Training a parameterized model largely depends on the landscape of the underlying loss function. We show that these bounds, as well as the variance of the loss itself, can be estimated efficiently and classically-versa--providing practical tools to study the loss landscapes of VQA models. This insight has direct implications for hybrid Quantum Generative Adrial Networks (qGANs), a generative model that can be reformulated as a VQA with an observable composed of local and global terms.
  arXiv  Detail & Related papers  (2023-09-22T07:38:13Z)
• Towards More Robust Interpretation via Local Gradient Alignment [37.464250451280336]
  We show that for every non-negative homogeneous neural network, a naive $ell$-robust criterion for gradients is textitnot normalization invariant. We propose to combine both $ell$ and cosine distance-based criteria as regularization terms to leverage the advantages of both in aligning the local gradient. We experimentally show that models trained with our method produce much more robust interpretations on CIFAR-10 and ImageNet-100.
  arXiv  Detail & Related papers  (2022-11-29T03:38:28Z)
• To be or not to be stable, that is the question: understanding neural networks for inverse problems [0.0]
  In this paper, we theoretically analyze the trade-off between stability and accuracy of neural networks. We propose different supervised and unsupervised solutions to increase the network stability and maintain a good accuracy.
  arXiv  Detail & Related papers  (2022-11-24T16:16:40Z)
• Dynamic Neural Diversification: Path to Computationally Sustainable Neural Networks [68.8204255655161]
  Small neural networks with a constrained number of trainable parameters, can be suitable resource-efficient candidates for many simple tasks. We explore the diversity of the neurons within the hidden layer during the learning process. We analyze how the diversity of the neurons affects predictions of the model.
  arXiv  Detail & Related papers  (2021-09-20T15:12:16Z)
• Better Training using Weight-Constrained Stochastic Dynamics [0.0]
  We employ constraints to control the parameter space of deep neural networks throughout training. The use of customized, appropriately designed constraints can reduce the vanishing/exploding problem. We provide a general approach to efficiently incorporate constraints into a gradient Langevin framework.
  arXiv  Detail & Related papers  (2021-06-20T14:41:06Z)
• Deep learning: a statistical viewpoint [120.94133818355645]
  Deep learning has revealed some major surprises from a theoretical perspective. In particular, simple gradient methods easily find near-perfect solutions to non-optimal training problems. We conjecture that specific principles underlie these phenomena.
  arXiv  Detail & Related papers  (2021-03-16T16:26:36Z)
• Non-Singular Adversarial Robustness of Neural Networks [58.731070632586594]
  Adrial robustness has become an emerging challenge for neural network owing to its over-sensitivity to small input perturbations. We formalize the notion of non-singular adversarial robustness for neural networks through the lens of joint perturbations to data inputs as well as model weights.
  arXiv  Detail & Related papers  (2021-02-23T20:59:30Z)
• Attribute-Guided Adversarial Training for Robustness to Natural Perturbations [64.35805267250682]
  We propose an adversarial training approach which learns to generate new samples so as to maximize exposure of the classifier to the attributes-space. Our approach enables deep neural networks to be robust against a wide range of naturally occurring perturbations.
  arXiv  Detail & Related papers  (2020-12-03T10:17:30Z)
• Gradient Starvation: A Learning Proclivity in Neural Networks [97.02382916372594]
  Gradient Starvation arises when cross-entropy loss is minimized by capturing only a subset of features relevant for the task. This work provides a theoretical explanation for the emergence of such feature imbalance in neural networks.
  arXiv  Detail & Related papers  (2020-11-18T18:52:08Z)
• QuantNet: Learning to Quantize by Learning within Fully Differentiable Framework [32.465949985191635]
  This paper proposes a meta-based quantizer named QuantNet, which utilizes a differentiable sub-network to directly binarize the full-precision weights. Our method not only solves the problem of gradient mismatching, but also reduces the impact of discretization errors, caused by the binarizing operation in the deployment.
  arXiv  Detail & Related papers  (2020-09-10T01:41:05Z)
• Implicit Bias in Deep Linear Classification: Initialization Scale vs Training Accuracy [71.25689267025244]
  We show how the transition is controlled by the relationship between the scale and how accurately we minimize the training loss. Our results indicate that some limit behaviors of gradient descent only kick in at ridiculous training accuracies.
  arXiv  Detail & Related papers  (2020-07-13T23:49:53Z)
• Constraint-Based Regularization of Neural Networks [0.0]
  We propose a method for efficiently incorporating constraints into a gradient Langevin framework for the training of deep neural networks. Appropriately designed, they reduce the vanishing/exploding gradient problem, control weight magnitudes and stabilize deep neural networks.
  arXiv  Detail & Related papers  (2020-06-17T19:28:41Z)
• 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)
• Path Sample-Analytic Gradient Estimators for Stochastic Binary Networks [78.76880041670904]
  In neural networks with binary activations and or binary weights the training by gradient descent is complicated. We propose a new method for this estimation problem combining sampling and analytic approximation steps. We experimentally show higher accuracy in gradient estimation and demonstrate a more stable and better performing training in deep convolutional models.
  arXiv  Detail & Related papers  (2020-06-04T21:51:21Z)
• Feature Purification: How Adversarial Training Performs Robust Deep Learning [66.05472746340142]
  We show a principle that we call Feature Purification, where we show one of the causes of the existence of adversarial examples is the accumulation of certain small dense mixtures in the hidden weights during the training process of a neural network. We present both experiments on the CIFAR-10 dataset to illustrate this principle, and a theoretical result proving that for certain natural classification tasks, training a two-layer neural network with ReLU activation using randomly gradient descent indeed this principle.
  arXiv  Detail & Related papers  (2020-05-20T16:56:08Z)
• Gradient $\ell_1$ Regularization for Quantization Robustness [70.39776106458858]
  We derive a simple regularization scheme that improves robustness against post-training quantization. By training quantization-ready networks, our approach enables storing a single set of weights that can be quantized on-demand to different bit-widths.
  arXiv  Detail & Related papers  (2020-02-18T12:31:34Z)
• Frosting Weights for Better Continual Training [22.554993259239307]
  Training a neural network model can be a lifelong learning process and is a computationally intensive one. Deep neural network models can suffer from catastrophic forgetting during retraining on new data. We propose two generic ensemble approaches, gradient boosting and meta-learning, to solve the problem.
  arXiv  Detail & Related papers  (2020-01-07T00:53:46Z)

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.