Dataset Distillation Efficiently Encodes Low-Dimensional Representations from Gradient-Based Learning of Non-Linear Tasks

• URL: http://arxiv.org/abs/2603.14830v1
• Date: Mon, 16 Mar 2026 05:14:34 GMT
• Title: Dataset Distillation Efficiently Encodes Low-Dimensional Representations from Gradient-Based Learning of Non-Linear Tasks
• Abstract summary: We analyze practical algorithms of dataset distillation applied to the gradient-based training of two-layer neural networks with width $L$.<n>By focusing on a non-linear task structure called multi-index model, we prove that the low-dimensional structure of the problem is efficiently encoded into the resulting distilled data.<n>This dataset reproduces a model with high ability for a required memory complexity of $tilde$$(r2d+L)$, where $d$ and $r$ are the input and intrinsic dimensions of the task.
• Score: 8.972144370022841
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Dataset distillation, a training-aware data compression technique, has recently attracted increasing attention as an effective tool for mitigating costs of optimization and data storage. However, progress remains largely empirical. Mechanisms underlying the extraction of task-relevant information from the training process and the efficient encoding of such information into synthetic data points remain elusive. In this paper, we theoretically analyze practical algorithms of dataset distillation applied to the gradient-based training of two-layer neural networks with width $L$. By focusing on a non-linear task structure called multi-index model, we prove that the low-dimensional structure of the problem is efficiently encoded into the resulting distilled data. This dataset reproduces a model with high generalization ability for a required memory complexity of $\tildeΘ$$(r^2d+L)$, where $d$ and $r$ are the input and intrinsic dimensions of the task. To the best of our knowledge, this is one of the first theoretical works that include a specific task structure, leverage its intrinsic dimensionality to quantify the compression rate and study dataset distillation implemented solely via gradient-based algorithms.

Related papers

• SAS: Semantic-aware Sampling for Generative Dataset Distillation [55.27114962330541]
  We introduce a semantic-aware perspective for dataset distillation by leveraging Contrastive Language-Image Pretraining (CLIP) as a semantic prior for post-sampling.<n>Our goal is to obtain distilled datasets that are not only compact but also semantically class-discriminative and diverse.
  arXiv  Detail & Related papers  (2026-05-18T08:05:46Z)
• Understanding Data Influence with Differential Approximation [63.817689230826595]
  We introduce a new formulation to approximate a sample's influence by accumulating the differences in influence between consecutive learning steps, which we term Diff-In.<n>By employing second-order approximations, we approximate these difference terms with high accuracy while eliminating the need for model convexity required by existing methods.<n>Our theoretical analysis demonstrates that Diff-In achieves significantly lower approximation error compared to existing influence estimators.
  arXiv  Detail & Related papers  (2025-08-20T11:59:32Z)
• Dataset Distillation as Pushforward Optimal Quantization [2.5892916589735457]
  We propose a synthetic training set that achieves similar performance to training on real data, with orders of magnitude less computational requirements.<n>In particular, we link existing disentangled dataset distillation methods to the classical optimal quantization and Wasserstein barycenter problems.<n>We achieve better performance and inter-model generalization on the ImageNet-1K dataset with trivial additional computation, and SOTA performance in higher image-per-class settings.
  arXiv  Detail & Related papers  (2025-01-13T20:41:52Z)
• Going Beyond Feature Similarity: Effective Dataset Distillation based on Class-Aware Conditional Mutual Information [43.44508080585033]
  We introduce conditional mutual information (CMI) to assess the class-aware complexity of a dataset.<n>We minimize the distillation loss while constraining the class-aware complexity of the synthetic dataset.
  arXiv  Detail & Related papers  (2024-12-13T08:10:47Z)
• Koopcon: A new approach towards smarter and less complex learning [13.053285552524052]
  In the era of big data, the sheer volume and complexity of datasets pose significant challenges in machine learning. This paper introduces an innovative Autoencoder-based dataset condensation model backed by Koopman operator theory. Inspired by the predictive coding mechanisms of the human brain, our model leverages a novel approach to encode and reconstruct data.
  arXiv  Detail & Related papers  (2024-05-22T17:47:14Z)
• Dataset Quantization [72.61936019738076]
  We present dataset quantization (DQ), a new framework to compress large-scale datasets into small subsets. DQ is the first method that can successfully distill large-scale datasets such as ImageNet-1k with a state-of-the-art compression ratio.
  arXiv  Detail & Related papers  (2023-08-21T07:24:29Z)
• Distill Gold from Massive Ores: Bi-level Data Pruning towards Efficient Dataset Distillation [96.92250565207017]
  We study the data efficiency and selection for the dataset distillation task. By re-formulating the dynamics of distillation, we provide insight into the inherent redundancy in the real dataset. We find the most contributing samples based on their causal effects on the distillation.
  arXiv  Detail & Related papers  (2023-05-28T06:53:41Z)
• Minimizing the Accumulated Trajectory Error to Improve Dataset Distillation [151.70234052015948]
  We propose a novel approach that encourages the optimization algorithm to seek a flat trajectory. We show that the weights trained on synthetic data are robust against the accumulated errors perturbations with the regularization towards the flat trajectory. Our method, called Flat Trajectory Distillation (FTD), is shown to boost the performance of gradient-matching methods by up to 4.7%.
  arXiv  Detail & Related papers  (2022-11-20T15:49:11Z)
• Dataset Distillation using Neural Feature Regression [32.53291298089172]
  We develop an algorithm for dataset distillation using neural Feature Regression with Pooling (FRePo) FRePo achieves state-of-the-art performance with an order of magnitude less memory requirement and two orders of magnitude faster training than previous methods. We show that high-quality distilled data can greatly improve various downstream applications, such as continual learning and membership inference defense.
  arXiv  Detail & Related papers  (2022-06-01T19:02:06Z)
• Learning to Generate Synthetic Training Data using Gradient Matching and Implicit Differentiation [77.34726150561087]
  This article explores various data distillation techniques that can reduce the amount of data required to successfully train deep networks. Inspired by recent ideas, we suggest new data distillation techniques based on generative teaching networks, gradient matching, and the Implicit Function Theorem.
  arXiv  Detail & Related papers  (2022-03-16T11:45:32Z)
• Unsupervised Learning of slow features for Data Efficient Regression [15.73372211126635]
  We propose the slow variational autoencoder (S-VAE), an extension to the $beta$-VAE which applies a temporal similarity constraint to the latent representations. We evaluate the three methods against their data-efficiency on down-stream tasks using a synthetic 2D ball tracking dataset, a dataset from a reinforcent learning environment and a dataset generated using the DeepMind Lab environment.
  arXiv  Detail & Related papers  (2020-12-11T12:19:45Z)

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.