Saved in:
Bibliographic Details
Main Authors: Liu, Dai, Gu, Jindong, Cao, Hu, Trinitis, Carsten, Schulz, Martin
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.14245
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916330314137600
author Liu, Dai
Gu, Jindong
Cao, Hu
Trinitis, Carsten
Schulz, Martin
author_facet Liu, Dai
Gu, Jindong
Cao, Hu
Trinitis, Carsten
Schulz, Martin
contents Dataset Distillation is used to create a concise, yet informative, synthetic dataset that can replace the original dataset for training purposes. Some leading methods in this domain prioritize long-range matching, involving the unrolling of training trajectories with a fixed number of steps (NS) on the synthetic dataset to align with various expert training trajectories. However, traditional long-range matching methods possess an overfitting-like problem, the fixed step size NS forces synthetic dataset to distortedly conform seen expert training trajectories, resulting in a loss of generality-especially to those from unencountered architecture. We refer to this as the Accumulated Mismatching Problem (AMP), and propose a new approach, Automatic Training Trajectories (ATT), which dynamically and adaptively adjusts trajectory length NS to address the AMP. Our method outperforms existing methods particularly in tests involving cross-architectures. Moreover, owing to its adaptive nature, it exhibits enhanced stability in the face of parameter variations.
format Preprint
id arxiv_https___arxiv_org_abs_2407_14245
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Dataset Distillation by Automatic Training Trajectories
Liu, Dai
Gu, Jindong
Cao, Hu
Trinitis, Carsten
Schulz, Martin
Computer Vision and Pattern Recognition
Dataset Distillation is used to create a concise, yet informative, synthetic dataset that can replace the original dataset for training purposes. Some leading methods in this domain prioritize long-range matching, involving the unrolling of training trajectories with a fixed number of steps (NS) on the synthetic dataset to align with various expert training trajectories. However, traditional long-range matching methods possess an overfitting-like problem, the fixed step size NS forces synthetic dataset to distortedly conform seen expert training trajectories, resulting in a loss of generality-especially to those from unencountered architecture. We refer to this as the Accumulated Mismatching Problem (AMP), and propose a new approach, Automatic Training Trajectories (ATT), which dynamically and adaptively adjusts trajectory length NS to address the AMP. Our method outperforms existing methods particularly in tests involving cross-architectures. Moreover, owing to its adaptive nature, it exhibits enhanced stability in the face of parameter variations.
title Dataset Distillation by Automatic Training Trajectories
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.14245