Xiuyuan Cheng

Conference Proceedings of Machine Learning Research · January 1, 2025 Computing optimal transport (OT) for general high-dimensional data has been a longstanding challenge. Despite much progress, most of the efforts including neural network methods have been focused on the static formulation of the OT problem. The current wor ... Cite

Consistency Posterior Sampling for Diverse Image Synthesis

Conference Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition · January 1, 2025 Posterior sampling in high-dimensional spaces using generative models holds significant promise for various applications, including but not limited to inverse problems and guided generation tasks. Generating diverse posterior samples remains expensive, as ... Full text Cite

STAGE-REGULARIZED NEURAL STEIN CRITICS FOR TESTING GOODNESS-OF-FIT OF GENERATIVE MODELS

Conference ICASSP IEEE International Conference on Acoustics Speech and Signal Processing Proceedings · January 1, 2024 Learning to differentiate model distributions from observed data is a fundamental problem in statistics and machine learning, and high-dimensional data remains a challenging setting for such problems. Metrics that quantify the disparity in probability dist ... Full text Cite

Training Neural Networks for Sequential Change-Point Detection

Conference ICASSP IEEE International Conference on Acoustics Speech and Signal Processing Proceedings · January 1, 2023 Detecting an abrupt distributional shift of a data stream, known as change-point detection, is a fundamental problem in statistics and machine learning. We introduce a novel approach for online change-point detection using neural net-works. To be specific, ... Full text Cite

NEURAL SPECTRAL MARKED POINT PROCESSES

Conference Iclr 2022 10th International Conference on Learning Representations · January 1, 2022 Self- and mutually-exciting point processes are popular models in machine learning and statistics for dependent discrete event data. To date, most existing models assume stationary kernels (including the classical Hawkes processes) and simple parametric mo ... Cite

SpecNet2: Orthogonalization-free Spectral Embedding by Neural Networks

Conference Proceedings of Machine Learning Research · January 1, 2022 Spectral methods which represent data points by eigenvectors of kernel matrices or graph Laplacian matrices have been a primary tool in unsupervised data analysis. In many application scenarios, parametrizing the spectral embedding by a neural network that ... Cite

Convergence of Gaussian-smoothed optimal transport distance with sub-gamma distributions and dependent samples

Conference Proceedings of Machine Learning Research · January 1, 2021 The Gaussian-smoothed optimal transport (GOT) framework, recently proposed by Goldfeld et al., scales to high dimensions in estimation and provides an alternative to entropy regularization. This paper provides convergence guarantees for estimating the GOT ... Cite

Spatiotemporal Joint Filter Decomposition in 3D Convolutional Neural Networks

Conference Advances in Neural Information Processing Systems · January 1, 2021 In this paper, we introduce spatiotemporal joint filter decomposition to decouple spatial and temporal learning, while preserving spatiotemporal dependency in a video. A 3D convolutional filter is now jointly decomposed over a set of spatial and temporal f ... Cite

Neural Tangent Kernel Maximum Mean Discrepancy

Conference Advances in Neural Information Processing Systems · January 1, 2021 We present a novel neural network Maximum Mean Discrepancy (MMD) statistic by identifying a new connection between neural tangent kernel (NTK) and MMD. This connection enables us to develop a computationally efficient and memory-efficient approach to compu ... Cite

Graph Convolution with Low-rank Learn-able Local Filters

Conference Iclr 2021 9th International Conference on Learning Representations · January 1, 2021 Geometric variations like rotation, scaling, and viewpoint changes pose a significant challenge to visual understanding. One common solution is to directly model certain intrinsic structures, e.g., using landmarks. However, it then becomes non-trivial to b ... Cite

STOCHASTIC CONDITIONAL GENERATIVE NETWORKS WITH BASIS DECOMPOSITION

Conference 8th International Conference on Learning Representations Iclr 2020 · January 1, 2020 While generative adversarial networks (GANs) have revolutionized machine learning, a number of open questions remain to fully understand them and exploit their power. One of these questions is how to efficiently achieve proper diversity and sampling of the ... Cite

Variational Diffusion Autoencoders with Random Walk Sampling

Conference Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics · January 1, 2020 Variational autoencoders (VAEs) and generative adversarial networks (GANs) enjoy an intuitive connection to manifold learning: in training the decoder/generator is optimized to approximate a homeomorphism between the data distribution and the sampling spac ... Full text Cite

Butterfly-Net2: Simplified Butterfly-Net and Fourier Transform Initialization

Conference Proceedings of Machine Learning Research · January 1, 2020 Structured CNN designed using the prior information of problems potentially improves efficiency over conventional CNNs in various tasks in solving PDEs and inverse problems in signal processing. This paper introduces BNet2, a simplified Butterfly-Net and i ... Cite

RotDCF: Decomposition of convolutional filters for rotation-equivariant deep networks

Conference · May 6, 2019 Link to item Cite

Provable estimation of the number of blocks in block models

Conference Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics (AISTATS'18) · April 9, 2018 Link to item Cite

DCFNet: Deep Neural Network with Decomposed Convolutional Filters.

Conference ICML · 2018 Cite

DCFNet: Deep Neural Network with Decomposed Convolutional Filters

Conference Proceedings of Machine Learning Research · January 1, 2018 Filters in a Convolutional Neural Network (CNN) contain model parameters learned from enormous amounts of data. In this paper, we suggest to decompose convolutional filters in CNN as a truncated expansion with pre-fixed bases, namely the Decomposed Convolu ... Cite

Provable Estimation of the Number of Blocks in Block Models

Conference Proceedings of Machine Learning Research · January 1, 2018 Community detection is a fundamental unsupervised learning problem for unlabeled networks which has a broad range of applications. Many community detection algorithms assume that the number of clusters r is known apriori. In this paper, we propose an appro ... Cite