G-TRACER: Expected Sharpness Optimization

• URL: http://arxiv.org/abs/2306.13914v1
• Date: Sat, 24 Jun 2023 09:28:49 GMT
• Title: G-TRACER: Expected Sharpness Optimization
• Authors: John Williams, Stephen Roberts
• Abstract summary: G-TRACER promotes generalization by seeking flat minima, and has a sound theoretical basis as an approximation to a natural-gradient descent based optimization of a generalized Bayes objective. We show that the method converges to a neighborhood of a local minimum of the unregularized objective, and demonstrate competitive performance on a number of benchmark computer vision and NLP datasets.
• Score: 1.2183405753834562
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We propose a new regularization scheme for the optimization of deep learning architectures, G-TRACER ("Geometric TRACE Ratio"), which promotes generalization by seeking flat minima, and has a sound theoretical basis as an approximation to a natural-gradient descent based optimization of a generalized Bayes objective. By augmenting the loss function with a TRACER, curvature-regularized optimizers (eg SGD-TRACER and Adam-TRACER) are simple to implement as modifications to existing optimizers and don't require extensive tuning. We show that the method converges to a neighborhood (depending on the regularization strength) of a local minimum of the unregularized objective, and demonstrate competitive performance on a number of benchmark computer vision and NLP datasets, with a particular focus on challenging low signal-to-noise ratio problems.

Related papers

• Fast Graph Sharpness-Aware Minimization for Enhancing and Accelerating Few-Shot Node Classification [53.727688136434345]
  Graph Neural Networks (GNNs) have shown superior performance in node classification. We present Fast Graph Sharpness-Aware Minimization (FGSAM) that integrates the rapid training of Multi-Layer Perceptrons with the superior performance of GNNs. Our proposed algorithm outperforms the standard SAM with lower computational costs in FSNC tasks.
  arXiv  Detail & Related papers  (2024-10-22T09:33:29Z)
• Gradient Normalization with(out) Clipping Ensures Convergence of Nonconvex SGD under Heavy-Tailed Noise with Improved Results [60.92029979853314]
  This paper investigates Gradient Normalization without (NSGDC) its gradient reduction variant (NSGDC-VR) We present significant improvements in the theoretical results for both algorithms.
  arXiv  Detail & Related papers  (2024-10-21T22:40:42Z)
• Alternating Minimization Schemes for Computing Rate-Distortion-Perception Functions with $f$-Divergence Perception Constraints [10.564071872770146]
  We study the computation of the rate-distortion-perception function (RDPF) for discrete memoryless sources. We characterize the optimal parametric solutions. We provide sufficient conditions on the distortion and the perception constraints.
  arXiv  Detail & Related papers  (2024-08-27T12:50:12Z)
• Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizer [52.09480867526656]
  We identify the source of misalignment as a form of distributional shift and uncertainty in learning human preferences. To mitigate overoptimization, we first propose a theoretical algorithm that chooses the best policy for an adversarially chosen reward model. Using the equivalence between reward models and the corresponding optimal policy, the algorithm features a simple objective that combines a preference optimization loss and a supervised learning loss.
  arXiv  Detail & Related papers  (2024-05-26T05:38:50Z)
• Beyond Single-Model Views for Deep Learning: Optimization versus Generalizability of Stochastic Optimization Algorithms [13.134564730161983]
  This paper adopts a novel approach to deep learning optimization, focusing on gradient descent (SGD) and its variants. We show that SGD and its variants demonstrate performance on par with flat-minimas like SAM, albeit with half the gradient evaluations. Our study uncovers several key findings regarding the relationship between training loss and hold-out accuracy, as well as the comparable performance of SGD and noise-enabled variants.
  arXiv  Detail & Related papers  (2024-03-01T14:55:22Z)
• Optimal Guarantees for Algorithmic Reproducibility and Gradient Complexity in Convex Optimization [55.115992622028685]
  Previous work suggests that first-order methods would need to trade-off convergence rate (gradient convergence rate) for better. We demonstrate that both optimal complexity and near-optimal convergence guarantees can be achieved for smooth convex minimization and smooth convex-concave minimax problems.
  arXiv  Detail & Related papers  (2023-10-26T19:56:52Z)
• SING: A Plug-and-Play DNN Learning Technique [25.563053353709627]
  We propose SING (StabIlized and Normalized Gradient), a plug-and-play technique that improves the stability and robustness of Adam(W) SING is straightforward to implement and has minimal computational overhead, requiring only a layer-wise standardization of gradients fed to Adam(W)
  arXiv  Detail & Related papers  (2023-05-25T12:39:45Z)
• Dynamic Regularized Sharpness Aware Minimization in Federated Learning: Approaching Global Consistency and Smooth Landscape [59.841889495864386]
  In federated learning (FL), a cluster of local clients are chaired under the coordination of a global server. Clients are prone to overfit into their own optima, which extremely deviates from the global objective. ttfamily FedSMOO adopts a dynamic regularizer to guarantee the local optima towards the global objective. Our theoretical analysis indicates that ttfamily FedSMOO achieves fast $mathcalO (1/T)$ convergence rate with low bound generalization.
  arXiv  Detail & Related papers  (2023-05-19T10:47:44Z)
• BAMSProd: A Step towards Generalizing the Adaptive Optimization Methods to Deep Binary Model [34.093978443640616]
  Recent methods have significantly reduced the performance of Binary Neural Networks (BNNs) guaranteeing the effective and efficient training of BNNs is an unsolved problem. We propose a BAMSProd algorithm with a key observation that the convergence property of optimizing deep binary model is strongly related to the quantization errors.
  arXiv  Detail & Related papers  (2020-09-29T06:12:32Z)
• 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)

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.