arXiv cs.AI· Zhiqiang He, Zhi Liu·· 1 天前AI 评分22
Silent Neuron 理论：用 ReSiN 保持深度强化学习在自适应视频流中的可塑性
Silent Neuron Theory and Plasticity Preservation for Deep Reinforcement Learning in Adaptive Video Streaming
AI 导读
针对深度强化学习在自适应视频流中因神经网络可塑性丧失而难以适应异构带宽的问题，研究者提出 Silent Neuron 理论，指出现有 dormant neuron 指标不足以刻画可塑性下降，并据此设计 ReSiN，通过结合前向与反向传播状态重置神经元来保持可塑性。
正文 · 原文
Abstract:Adaptive video streaming optimizes Quality of Experience (QoE) metrics by selecting appropriate bitrates according to varying network bandwidth and user demands. In practice, however, real-world network bandwidth often exhibits heterogeneity relative to training environments. Current methods predominantly tackle this problem through learning-based approaches designed to improve generalization performance. Our systematic investigation reveals a critical limitation: neural networks suffer from plasticity loss, significantly impeding their ability to adapt to heterogeneous network conditions. Through theoretical analysis of neural propagation mechanisms, we demonstrate that existing dormant neuron metrics inadequately characterize neural plasticity loss. To address this limitation, we have developed the Silent Neuron theory, which provides a more comprehensive framework for understanding plasticity degradation. Based on these theoretical insights, we propose the Reset Silent Neuron (ReSiN), which preserves neural plasticity through strategic neuron resets guided by both forward and backward propagation states. Moreover, we establish a tighter performance bound for ReSiN under non-stationary network conditions. In our implementation of an adaptive video streaming system, ReSiN shows substantial gains under non-stationary conditions in which the available bandwidth alternates between high and low regimes, improving the QoE by 178%, and it remains beneficial under stationary real-world traces, where it improves the QoE by 1.6%. Code will be available: this https URL
| Comments: | This paper has been accepted by IEEE Transactions on Mobile Computing |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2505.01584 [cs.LG] |
| (or arXiv:2505.01584v5 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2505.01584 arXiv-issued DOI via DataCite |
Submission history
From: Zhiqiang He [view email]
[v1]
Fri, 2 May 2025 21:03:03 UTC (10,839 KB)
[v2]
Tue, 6 May 2025 02:36:10 UTC (10,839 KB)
[v3]
Thu, 14 May 2026 04:29:35 UTC (19,404 KB)
[v4]
Wed, 8 Jul 2026 10:04:35 UTC (18,806 KB)
[v5]
Wed, 11 Nov 2026 07:00:26 UTC (18,846 KB)
来源：arXiv cs.AI · arxiv.org