Search Results for author: Weiyu Xu

Found 27 papers, 1 papers with code

gcDLSeg: Integrating Graph-cut into Deep Learning for Binary Semantic Segmentation

no code implementations7 Dec 2023 Hui Xie, Weiyu Xu, Ya Xing Wang, John Buatti, Xiaodong Wu

To combine the strengths of both approaches, we propose in this study to integrate the graph-cut approach into a deep learning network for end-to-end learning.

Segmentation Semantic Segmentation

Optimal Cost Constrained Adversarial Attacks For Multiple Agent Systems

no code implementations1 Nov 2023 Ziqing Lu, Guanlin Liu, Lifeng Cai, Weiyu Xu

Finding optimal adversarial attack strategies is an important topic in reinforcement learning and the Markov decision process.

Adversarial Attack

Trust, but Verify: Robust Image Segmentation using Deep Learning

no code implementations25 Oct 2023 Fahim Ahmed Zaman, Xiaodong Wu, Weiyu Xu, Milan Sonka, Raghuraman Mudumbai

We describe a method for verifying the output of a deep neural network for medical image segmentation that is robust to several classes of random as well as worst-case perturbations i. e. adversarial attacks.

Image Segmentation Medical Image Segmentation +2

Outlier Detection Using Generative Models with Theoretical Performance Guarantees

no code implementations16 Oct 2023 Jirong Yi, Jingchao Gao, Tianming Wang, Xiaodong Wu, Weiyu Xu

We propose an outlier detection approach for reconstructing the ground-truth signals modeled by generative models under sparse outliers.

Outlier Detection

Linear Progressive Coding for Semantic Communication using Deep Neural Networks

no code implementations27 Sep 2023 Eva Riherd, Raghu Mudumbai, Weiyu Xu

We propose a general method for semantic representation of images and other data using progressive coding.

Image Compression

Distributed Dual Coordinate Ascent with Imbalanced Data on a General Tree Network

no code implementations28 Aug 2023 Myung Cho, Lifeng Lai, Weiyu Xu

In this paper, we investigate the impact of imbalanced data on the convergence of distributed dual coordinate ascent in a tree network for solving an empirical loss minimization problem in distributed machine learning.

To AI or not to AI, to Buy Local or not to Buy Local: A Mathematical Theory of Real Price

no code implementations9 May 2023 Huan Cai, Catherine Xu, Weiyu Xu

In the past several decades, the world's economy has become increasingly globalized.

Optimal Compression for Minimizing Classification Error Probability: an Information-Theoretic Approach

no code implementations3 Nov 2022 Jingchao Gao, Ao Tang, Weiyu Xu

We then provide analytical and computational methods to characterize the optimal trade-off between data compression and classification error probability.

Classification Data Compression

A deep learning network with differentiable dynamic programming for retina OCT surface segmentation

no code implementations8 Oct 2022 Hui Xie, Weiyu Xu, Xiaodong Wu

Unfortunately, due to the scarcity of training data in medical imaging, it is challenging for DL networks to learn the global structure of the target surfaces, including surface smoothness.

Model Optimization Segmentation

Optimal Pooling Matrix Design for Group Testing with Dilution (Row Degree) Constraints

no code implementations5 Aug 2020 Jirong Yi, Myung Cho, Xiaodong Wu, Raghu Mudumbai, Weiyu Xu

In this paper, we consider the problem of designing optimal pooling matrix for group testing (for example, for COVID-19 virus testing) with the constraint that no more than $r>0$ samples can be pooled together, which we call "dilution constraint".

Derivation of Information-Theoretically Optimal Adversarial Attacks with Applications to Robust Machine Learning

no code implementations28 Jul 2020 Jirong Yi, Raghu Mudumbai, Weiyu Xu

We consider the theoretical problem of designing an optimal adversarial attack on a decision system that maximally degrades the achievable performance of the system as measured by the mutual information between the degraded signal and the label of interest.

Adversarial Attack BIG-bench Machine Learning +1

Trust but Verify: An Information-Theoretic Explanation for the Adversarial Fragility of Machine Learning Systems, and a General Defense against Adversarial Attacks

no code implementations25 May 2019 Jirong Yi, Hui Xie, Leixin Zhou, Xiaodong Wu, Weiyu Xu, Raghuraman Mudumbai

In this paper, we present a simple hypothesis about a feature compression property of artificial intelligence (AI) classifiers and present theoretical arguments to show that this hypothesis successfully accounts for the observed fragility of AI classifiers to small adversarial perturbations.

Feature Compression

Fast Single Image Reflection Suppression via Convex Optimization

1 code implementation CVPR 2019 Yang Yang, Wenye Ma, Yin Zheng, Jian-Feng Cai, Weiyu Xu

Removing undesired reflections from images taken through the glass is of great importance in computer vision.

BIG-bench Machine Learning

An Information-Theoretic Explanation for the Adversarial Fragility of AI Classifiers

no code implementations27 Jan 2019 Hui Xie, Jirong Yi, Weiyu Xu, Raghu Mudumbai

We present a simple hypothesis about a compression property of artificial intelligence (AI) classifiers and present theoretical arguments to show that this hypothesis successfully accounts for the observed fragility of AI classifiers to small adversarial perturbations.

Outlier Detection using Generative Models with Theoretical Performance Guarantees

no code implementations26 Oct 2018 Jirong Yi, Anh Duc Le, Tianming Wang, Xiaodong Wu, Weiyu Xu

In this paper, we propose a generative model neural network approach for reconstructing the ground truth signals under sparse outliers.

Outlier Detection

Necessary and Sufficient Null Space Condition for Nuclear Norm Minimization in Low-Rank Matrix Recovery

no code implementations14 Feb 2018 Jirong Yi, Weiyu Xu

In [12, 14, 15], the authors established the necessary and sufficient null space conditions for nuclear norm minimization to recover every possible low-rank matrix with rank at most r (the strong null space condition).

Collaborative Filtering

Separation-Free Super-Resolution from Compressed Measurements is Possible: an Orthonormal Atomic Norm Minimization Approach

no code implementations4 Nov 2017 Weiyu Xu, Jirong Yi, Soura Dasgupta, Jian-Feng Cai, Mathews Jacob, Myung Cho

However, it is known that in order for TV minimization and atomic norm minimization to recover the missing data or the frequencies, the underlying $R$ frequencies are required to be well-separated, even when the measurements are noiseless.

Super-Resolution

Distributed Dual Coordinate Ascent in General Tree Networks and Communication Network Effect on Synchronous Machine Learning

no code implementations14 Mar 2017 Myung Cho, Lifeng Lai, Weiyu Xu

Additionally, we show that adapting number of local and global iterations to network communication delays in the distributed dual coordinated ascent algorithm can improve its convergence speed.

BIG-bench Machine Learning

Precise Phase Transition of Total Variation Minimization

no code implementations15 Sep 2015 Bingwen Zhang, Weiyu Xu, Jian-Feng Cai, Lifeng Lai

Characterizing the phase transitions of convex optimizations in recovering structured signals or data is of central importance in compressed sensing, machine learning and statistics.

Denoising

Projected Wirtinger Gradient Descent for Low-Rank Hankel Matrix Completion in Spectral Compressed Sensing

no code implementations14 Jul 2015 Jian-Feng Cai, Suhui Liu, Weiyu Xu

This paper considers reconstructing a spectrally sparse signal from a small number of randomly observed time-domain samples.

Matrix Completion

Robust recovery of complex exponential signals from random Gaussian projections via low rank Hankel matrix reconstruction

no code implementations10 Mar 2015 Jian-Feng Cai, Xiaobo Qu, Weiyu Xu, Gui-Bo Ye

Our method can be applied to spectral compressed sensing where the signal of interest is a superposition of $R$ complex sinusoids.

Precise Semidefinite Programming Formulation of Atomic Norm Minimization for Recovering d-Dimensional ($d\geq 2$) Off-the-Grid Frequencies

no code implementations2 Dec 2013 Weiyu Xu, Jian-Feng Cai, Kumar Vijay Mishra, Myung Cho, Anton Kruger

Recent research in off-the-grid compressed sensing (CS) has demonstrated that, under certain conditions, one can successfully recover a spectrally sparse signal from a few time-domain samples even though the dictionary is continuous.

Universally Elevating the Phase Transition Performance of Compressed Sensing: Non-Isometric Matrices are Not Necessarily Bad Matrices

no code implementations17 Jul 2013 Weiyu Xu, Myung Cho

In this paper, we show that a polynomial-time algorithm can universally elevate the phase-transition performance of compressed sensing, compared with $\ell_1$ minimization, even for signals with constant-modulus nonzero elements.

Precisely Verifying the Null Space Conditions in Compressed Sensing: A Sandwiching Algorithm

no code implementations11 Jun 2013 Myung Cho, Weiyu Xu

In this paper, we first propose a series of new polynomial-time algorithms to compute upper bounds on $\alpha_k$.

Guarantees of Total Variation Minimization for Signal Recovery

no code implementations28 Jan 2013 Jian-Feng Cai, Weiyu Xu

In this paper, we consider using total variation minimization to recover signals whose gradients have a sparse support, from a small number of measurements.

Cannot find the paper you are looking for? You can Submit a new open access paper.