Dataset Distillation by Matching Training Trajectories
This work addresses the problem of efficiently compressing datasets for machine learning, offering a novel approach that improves upon prior methods in terms of accuracy and scalability.
The paper tackles dataset distillation by proposing a method that optimizes synthetic data to match training trajectories of models trained on real data, achieving superior performance over existing methods and enabling distillation of higher-resolution visual data.
Dataset distillation is the task of synthesizing a small dataset such that a model trained on the synthetic set will match the test accuracy of the model trained on the full dataset. In this paper, we propose a new formulation that optimizes our distilled data to guide networks to a similar state as those trained on real data across many training steps. Given a network, we train it for several iterations on our distilled data and optimize the distilled data with respect to the distance between the synthetically trained parameters and the parameters trained on real data. To efficiently obtain the initial and target network parameters for large-scale datasets, we pre-compute and store training trajectories of expert networks trained on the real dataset. Our method handily outperforms existing methods and also allows us to distill higher-resolution visual data.