CVMar 11, 2025

Efficient Dataset Distillation through Low-Rank Space Sampling

arXiv:2503.07998v11 citationsh-index: 19ICASSP
Originality Incremental advance
AI Analysis

This work addresses the need for efficient training strategies in deep learning by reducing data redundancy, though it is incremental as it builds on existing dataset distillation methods.

The paper tackles the problem of dataset distillation by proposing a method that captures common features among data using low-rank approximations, achieving an average performance improvement of 9.9% over baselines on CIFAR-10, CIFAR-100, and SVHN datasets.

Huge amount of data is the key of the success of deep learning, however, redundant information impairs the generalization ability of the model and increases the burden of calculation. Dataset Distillation (DD) compresses the original dataset into a smaller but representative subset for high-quality data and efficient training strategies. Existing works for DD generate synthetic images by treating each image as an independent entity, thereby overlooking the common features among data. This paper proposes a dataset distillation method based on Matching Training Trajectories with Low-rank Space Sampling(MTT-LSS), which uses low-rank approximations to capture multiple low-dimensional manifold subspaces of the original data. The synthetic data is represented by basis vectors and shared dimension mappers from these subspaces, reducing the cost of generating individual data points while effectively minimizing information redundancy. The proposed method is tested on CIFAR-10, CIFAR-100, and SVHN datasets, and outperforms the baseline methods by an average of 9.9%.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes