基于动态模式的剪枝方法，让时序网络更轻更快。
Rethinking Reservoir Pruning: A Dynamical Perspective for Echo State Networks
- 通过轨迹平均雅可比格拉姆矩阵识别关键神经元
- 在混沌与真实时间序列上保持或提升预测精度
- 适合追求高效时序建模的开发者和研究者
回声状态网络（ESNs）为时序预测提供了高效框架，但其随机初始化的动态储备池常存在过度参数化和动态冗余问题。现有剪枝方法多依赖静态连接或激活统计，可能忽略影响输入驱动状态转移的关键神经元。本文提出动态模式剪枝（DMP），通过轨迹平均雅可比格拉姆矩阵评估神经元对主导转移模式的贡献，并据此排序剔除低影响单元，仅重训输出层。在混沌及真实世界时间序列基准测试中，DMP在减少冗余储备池组件的同时保持或提升预测准确性。结果表明，动态影响是超越静态结构重要性的有效剪枝依据。
原文摘要 · Abstract (English)
Echo State Networks (ESNs) offer an efficient framework for temporal prediction, but their randomly initialized reservoirs are often over-parameterized and dynamically redundant. Existing pruning methods largely rely on static connectivity or activation statistics, which may overlook neurons that shape input-driven state transitions. We propose Dynamical Mode Pruning (DMP), a reservoir pruning method that ranks neurons by their contribution to dominant transition modes obtained from a trajectory-averaged Jacobian Gramian. DMP removes low-impact units and retrains only the readout. Experiments on chaotic and real-world time-series benchmarks show that DMP improves or preserves forecasting accuracy while reducing redundant reservoir components. Our results suggest that dynamical influence is a useful criterion for reservoir refinement beyond static structural importance alone.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。