Neural Spectral Decomposition for Dataset Distillation

• URL: http://arxiv.org/abs/2408.16236v1
• Date: Thu, 29 Aug 2024 03:26:14 GMT
• Title: Neural Spectral Decomposition for Dataset Distillation
• Authors: Shaolei Yang, Shen Cheng, Mingbo Hong, Haoqiang Fan, Xing Wei, Shuaicheng Liu,
• Abstract summary: We propose Neural Spectrum Decomposition, a generic decomposition framework for dataset distillation. We aim to discover the low-rank representation of the entire dataset and perform distillation efficiently. Our results demonstrate that our approach achieves state-of-the-art performance on benchmarks, including CIFAR10, CIFAR100, Tiny Imagenet, and ImageNet Subset.
• Score: 48.59372086450124
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: In this paper, we propose Neural Spectrum Decomposition, a generic decomposition framework for dataset distillation. Unlike previous methods, we consider the entire dataset as a high-dimensional observation that is low-rank across all dimensions. We aim to discover the low-rank representation of the entire dataset and perform distillation efficiently. Toward this end, we learn a set of spectrum tensors and transformation matrices, which, through simple matrix multiplication, reconstruct the data distribution. Specifically, a spectrum tensor can be mapped back to the image space by a transformation matrix, and efficient information sharing during the distillation learning process is achieved through pairwise combinations of different spectrum vectors and transformation matrices. Furthermore, we integrate a trajectory matching optimization method guided by a real distribution. Our experimental results demonstrate that our approach achieves state-of-the-art performance on benchmarks, including CIFAR10, CIFAR100, Tiny Imagenet, and ImageNet Subset. Our code are available at \url{https://github.com/slyang2021/NSD}.

Related papers

• Graph Generation via Spectral Diffusion [51.60814773299899]
  We present GRASP, a novel graph generative model based on 1) the spectral decomposition of the graph Laplacian matrix and 2) a diffusion process. Specifically, we propose to use a denoising model to sample eigenvectors and eigenvalues from which we can reconstruct the graph Laplacian and adjacency matrix. Our permutation invariant model can also handle node features by concatenating them to the eigenvectors of each node.
  arXiv  Detail & Related papers  (2024-02-29T09:26:46Z)
• Datacube segmentation via Deep Spectral Clustering [76.48544221010424]
  Extended Vision techniques often pose a challenge in their interpretation. The huge dimensionality of data cube spectra poses a complex task in its statistical interpretation. In this paper, we explore the possibility of applying unsupervised clustering methods in encoded space. A statistical dimensional reduction is performed by an ad hoc trained (Variational) AutoEncoder, while the clustering process is performed by a (learnable) iterative K-Means clustering algorithm.
  arXiv  Detail & Related papers  (2024-01-31T09:31:28Z)
• Dataset Distillation with Convexified Implicit Gradients [69.16247946639233]
  We show how implicit gradients can be effectively used to compute meta-gradient updates. We further equip the algorithm with a convexified approximation that corresponds to learning on top of a frozen finite-width neural kernel.
  arXiv  Detail & Related papers  (2023-02-13T23:53:16Z)
• Orthonormal Convolutions for the Rotation Based Iterative Gaussianization [64.44661342486434]
  This paper elaborates an extension of rotation-based iterative Gaussianization, RBIG, which makes image Gaussianization possible. In images its application has been restricted to small image patches or isolated pixels, because rotation in RBIG is based on principal or independent component analysis. We present the emphConvolutional RBIG: an extension that alleviates this issue by imposing that the rotation in RBIG is a convolution.
  arXiv  Detail & Related papers  (2022-06-08T12:56:34Z)
• SUMD: Super U-shaped Matrix Decomposition Convolutional neural network for Image denoising [0.0]
  We introduce the matrix decomposition module(MD) in the network to establish the global context feature. Inspired by the design of multi-stage progressive restoration of U-shaped architecture, we further integrate the MD module into the multi-branches. Our model(SUMD) can produce comparable visual quality and accuracy results with Transformer-based methods.
  arXiv  Detail & Related papers  (2022-04-11T04:38:34Z)
• Regularization by Denoising Sub-sampled Newton Method for Spectral CT Multi-Material Decomposition [78.37855832568569]
  We propose to solve a model-based maximum-a-posterior problem to reconstruct multi-materials images with application to spectral CT. In particular, we propose to solve a regularized optimization problem based on a plug-in image-denoising function. We show numerical and experimental results for spectral CT materials decomposition.
  arXiv  Detail & Related papers  (2021-03-25T15:20:10Z)
• LADMM-Net: An Unrolled Deep Network For Spectral Image Fusion From Compressive Data [6.230751621285322]
  Hyperspectral (HS) and multispectral (MS) image fusion aims at estimating a high-resolution spectral image from a low-spatial-resolution HS image and a low-spectral-resolution MS image. In this work, a deep learning architecture under the algorithm unrolling approach is proposed for solving the fusion problem from HS and MS compressive measurements.
  arXiv  Detail & Related papers  (2021-03-01T12:04:42Z)
• Hyperspectral Image Denoising with Partially Orthogonal Matrix Vector Tensor Factorization [42.56231647066719]
  Hyperspectral image (HSI) has some advantages over natural image for various applications due to the extra spectral information. During the acquisition, it is often contaminated by severe noises including Gaussian noise, impulse noise, deadlines, and stripes. We present a HSI restoration method named smooth and robust low rank tensor recovery.
  arXiv  Detail & Related papers  (2020-06-29T02:10:07Z)
• Deeply Learned Spectral Total Variation Decomposition [8.679020335206753]
  We present a neural network approximation of a non-linear spectral decomposition. We report up to four orders of magnitude ($times 10,000$) speedup in processing of mega-pixel size images.
  arXiv  Detail & Related papers  (2020-06-17T17:10:43Z)

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.