Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lyu, Jiacheng, Bao, Bihua, Yan, Shiyun
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2601.02081
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913003396399104
author Lyu, Jiacheng
Bao, Bihua
Yan, Shiyun
author_facet Lyu, Jiacheng
Bao, Bihua
Yan, Shiyun
contents Massive datasets often contain redundancy that inflates computational costs without improving generalization. Existing data reduction methods are typically task-agnostic, discarding informative boundary samples and yielding suboptimal performance. We propose Adversarial Soft-Selection Subsampling (ASSS), a differentiable framework that casts data reduction as a minimax game between a learnable selector and a task network. Using Gumbel-Softmax relaxation, ASSS enables end-to-end gradient flow and is theoretically grounded in the information bottleneck principle. Experiments on multiple benchmarks show that ASSS achieves a performance retention rate (PRR) of 98.9% while using only 30% of the data, significantly outperforming random sampling, K-means, and gradient-based methods. Visualizations confirm that ASSS preferentially retains samples near decision boundaries. The framework is scalable, fully differentiable, and easily integrated into existing training pipelines. This work introduces a new paradigm for task-aware data reduction that directly optimizes subset selection for the downstream objective, offering a principled and practical solution to the scalability challenges in modern deep learning.
format Preprint
id arxiv_https___arxiv_org_abs_2601_02081
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ASSS: A Differentiable Adversarial Framework for Task-Aware Data Reduction
Lyu, Jiacheng
Bao, Bihua
Yan, Shiyun
Machine Learning
Massive datasets often contain redundancy that inflates computational costs without improving generalization. Existing data reduction methods are typically task-agnostic, discarding informative boundary samples and yielding suboptimal performance. We propose Adversarial Soft-Selection Subsampling (ASSS), a differentiable framework that casts data reduction as a minimax game between a learnable selector and a task network. Using Gumbel-Softmax relaxation, ASSS enables end-to-end gradient flow and is theoretically grounded in the information bottleneck principle. Experiments on multiple benchmarks show that ASSS achieves a performance retention rate (PRR) of 98.9% while using only 30% of the data, significantly outperforming random sampling, K-means, and gradient-based methods. Visualizations confirm that ASSS preferentially retains samples near decision boundaries. The framework is scalable, fully differentiable, and easily integrated into existing training pipelines. This work introduces a new paradigm for task-aware data reduction that directly optimizes subset selection for the downstream objective, offering a principled and practical solution to the scalability challenges in modern deep learning.
title ASSS: A Differentiable Adversarial Framework for Task-Aware Data Reduction
topic Machine Learning
url https://arxiv.org/abs/2601.02081