一句话：提出 Ballot 剪枝框架，通过冲突检测子网络选择与改进训练提升 DNN 驱动软件的公平性。
随着 Software 3.0 时代到来，将大型模型压缩并集成到软件系统中成为趋势，但模型压缩会损害 DNN 驱动软件的公平性。本文聚焦彩票假设（LTH）剪枝，指出其公平性问题同时源于子网络选择与训练过程，且现有补救措施不足。为此，作者提出剪枝框架 Ballot：采用基于冲突检测的子网络选择来寻找准确且公平的子网络，并结合精炼训练过程获得高性能模型。在五个常用数据集和三个模型上评估，Ballot 相比 Magnitude Pruning、Standard LTH、SafeCompress 和 FairScratch，公平性分别提升 38.00%、33.91%、17.96% 和 35.82%。
With the emergence of the Software 3.0 era, there is a growing trend of compressing and integrating large models into software systems, with significant societal implications. Regrettably, in numerous instances, model compression techniques impact the fairness performance of these models and thus the ethical behavior of DNN-powered software. One of the most notable example is the Lottery Ticket Hypothesis (LTH), a prevailing model pruning approach. This paper demonstrates that fairness issue of LTH-based pruning arises from both its subnetwork selection and training procedures, highlighting the inadequacy of existing remedies. To address this, we propose a novel pruning framework, Ballot, which employs a novel conflict-detection-based subnetwork selection to find accurate and fair subnetworks, coupled with a refined training process to attain a high-performance model, thereby improving the fairness of DNN-powered software. By means of this procedure, Ballot improves the fairness of pruning by 38.00%, 33.91%, 17.96%, and 35.82% compared to state-of-the-art baselines, namely Magnitude Pruning, Standard LTH, SafeCompress, and FairScratch respectively, based on our evaluation of five popular datasets and three widely used models. Our code is available at https://anonymous.4open.science/r/Ballot-506E.
DOI 原文 ·
@inproceedings{paperbot8472,
title = {Efficient DNN-Powered Software with Fair Sparse Models},
author = {Xuanqi Gao and Weipeng Jiang and Juan Zhai and Shiqing Ma and Xiaoyu Zhang and Chao Shen},
booktitle = {ISSTA},
year = {2024},
doi = {10.1145/3650212.3680336}
}