Description
Based on the understanding that the flat local minima of the empirical risk cause the model to generalize better. Adversarial Model Perturbation (AMP) improves generalization via minimizing the AMP loss, which is obtained from the empirical risk by applying the worst norm-bounded perturbation on each point in the parameter space.
Papers Using This Method
AmpLyze: A Deep Learning Model for Predicting the Hemolytic Concentration2025-07-10GLAMP: An Approximate Message Passing Framework for Transfer Learning with Applications to Lasso-based Estimators2025-05-28Graph Neural Network Aided Detection for the Multi-User Multi-Dimensional Index Modulated Uplink2025-05-27Detailed Evaluation of Modern Machine Learning Approaches for Optic Plastics Sorting2025-05-22Aliasing Reduction in Neural Amp Modeling by Smoothing Activations2025-05-07Efficient LLMs with AMP: Attention Heads and MLP Pruning2025-04-29Targeted AMP generation through controlled diffusion with efficient embeddings2025-04-24Tensor Generalized Approximate Message Passing2025-03-25ZO2: Scalable Zeroth-Order Fine-Tuning for Extremely Large Language Models with Limited GPU Memory2025-03-16Asymptotics of Non-Convex Generalized Linear Models in High-Dimensions: A proof of the replica formula2025-02-27Noise-conditioned Energy-based Annealed Rewards (NEAR): A Generative Framework for Imitation Learning from Observation2025-01-24Comprehensive Assessment of BERT-Based Methods for Predicting Antimicrobial Peptides2024-12-24SGAC: A Graph Neural Network Framework for Imbalanced and Structure-Aware AMP Classification2024-12-20Unifying AMP Algorithms for Rotationally-Invariant Models2024-12-02Fast, robust approximate message passing2024-11-05Building Conformal Prediction Intervals with Approximate Message Passing2024-10-21Deep Unfolded Approximate Message Passing for Quantitative Acoustic Microscopy Image Reconstruction2024-09-20Unrolled denoising networks provably learn optimal Bayesian inference2024-09-19Digital Semantic Communications: An Alternating Multi-Phase Training Strategy with Mask Attack2024-08-09Distributed Memory Approximate Message Passing2024-07-25