Zobrazeno 1 - 10
of 31
pro vyhledávání: '"Pooladzandi, Omead"'
Data poisoning attacks pose a significant threat to the integrity of machine learning models by leading to misclassification of target distribution data by injecting adversarial examples during training. Existing state-of-the-art (SoTA) defense metho
Externí odkaz:
http://arxiv.org/abs/2405.19376
Train-time data poisoning attacks threaten machine learning models by introducing adversarial examples during training, leading to misclassification. Current defense methods often reduce generalization performance, are attack-specific, and impose sig
Externí odkaz:
http://arxiv.org/abs/2405.18627
Autor:
Pooladzandi, Omead, Li, Xi-Lin
We present a novel approach to accelerate stochastic gradient descent (SGD) by utilizing curvature information obtained from Hessian-vector products or finite differences of parameters and gradients, similar to the BFGS algorithm. Our approach involv
Externí odkaz:
http://arxiv.org/abs/2402.04553
We propose a composable framework for latent space image augmentation that allows for easy combination of multiple augmentations. Image augmentation has been shown to be an effective technique for improving the performance of a wide variety of image
Externí odkaz:
http://arxiv.org/abs/2303.03462
Generative models have the ability to synthesize data points drawn from the data distribution, however, not all generated samples are high quality. In this paper, we propose using a combination of coresets selection methods and ``entropic regularizat
Externí odkaz:
http://arxiv.org/abs/2302.00138
Autor:
Pooladzandi, Omead, Zhou, Yiming
We explore the usage of the Levenberg-Marquardt (LM) algorithm for regression (non-linear least squares) and classification (generalized Gauss-Newton methods) tasks in neural networks. We compare the performance of the LM method with other popular fi
Externí odkaz:
http://arxiv.org/abs/2212.08769
A vast amount of expert and domain knowledge is captured by causal structural priors, yet there has been little research on testing such priors for generalization and data synthesis purposes. We propose a novel model architecture, Causal Structural H
Externí odkaz:
http://arxiv.org/abs/2210.11275
Publikováno v:
International Conference on Machine Learning 2022
Training machine learning models on massive datasets incurs substantial computational costs. To alleviate such costs, there has been a sustained effort to develop data-efficient training methods that can carefully select subsets of the training examp
Externí odkaz:
http://arxiv.org/abs/2207.13887
Variational autoencoders (VAEs) and other generative methods have garnered growing interest not just for their generative properties but also for the ability to dis-entangle a low-dimensional latent variable space. However, few existing generative mo
Externí odkaz:
http://arxiv.org/abs/2207.01575
Imitation learning is the task of replicating expert policy from demonstrations, without access to a reward function. This task becomes particularly challenging when the expert exhibits a mixture of behaviors. Prior work has introduced latent variabl
Externí odkaz:
http://arxiv.org/abs/2205.03484