AI 中文总结
TSAE提出结构化稀疏自编码器，通过共享/私有字典、门控编码和时间连续性约束分解时间序列预测模型的隐藏状态，并配合TSEVAL协议审计，在三个数据集上以更低重建误差和NFRE实现可解释特征。
AI 中文摘要
时间序列预测为能源调度、工业运营和环境监测中的关键决策提供信息；理解模型所依赖的模式对于评估可靠性和识别故障至关重要。输入归因识别重要变量和时间段，但对内部特征的洞察有限，而标准稀疏自编码器（SAE）目标不直接约束跨变量结构或时间连续性。我们引入TSAE，一种用于预测表示的结构化稀疏自编码器，将隐藏状态分解为可单独检查的特征。TSAE通过共享和变量路由的私有字典组织跨变量结构，通过门控编码将特征检测与幅度估计分离，并根据原始片段相似性约束相邻稀疏编码的变化。这些机制支持对变量上下文、激活强度和时间演化的分析。预测一致性微调进一步改善了冻结预测器输出的保留。随附的TSEVAL协议分别审计保真度、特征连贯性和物理校准，以夯实特征解释。在冻结PatchTST对ETTh1、ETTh2和ETTm1的三种子实验中，TSAE在五个SAE中实现了最低的隐藏状态重建误差、归一化预测重建误差（NFRE）和特征转换率，且每令牌活动相当。NFRE相对于次优均值降低4.3%-26.4%。选择性、物理相关性和校准中的数据集相关权衡表明，保真度和时间稳定性增益需要独立的语义验证以支持特征解释。
英文摘要
Time-series forecasting informs critical decisions in energy dispatch, industrial operations, and environmental monitoring; understanding the patterns models rely on is essential for assessing reliability and identifying failures. Input attribution identifies important variables and time segments but offers limited insight into internal features, while standard sparse autoencoder (SAE) objectives do not directly constrain cross-variable structure or temporal continuity. We introduce TSAE, a structured sparse autoencoder for forecasting representations that decomposes hidden states into individually inspectable features. TSAE organizes cross-variable structure through shared and variable-routed private dictionaries, separates feature detection from magnitude estimation with gated encoding, and constrains neighboring sparse-code changes according to raw-segment similarity. These mechanisms support analysis of variable context, activation strength, and temporal evolution. Forecast-consistency fine-tuning further improves preservation of the frozen forecaster's outputs. The accompanying TSEVAL protocol separately audits fidelity, feature coherence, and physical calibration to ground feature interpretation. In three-seed experiments with frozen PatchTST on ETTh1, ETTh2, and ETTm1, TSAE achieves the lowest hidden-state reconstruction error, normalized forecast-reconstruction error (NFRE), and feature transition rate among five SAEs at comparable per-token activity. NFRE decreases by 4.3-26.4% relative to the next-best mean. Dataset-dependent tradeoffs in selectivity, physical correlation, and calibration show that fidelity and temporal-stability gains require independent semantic validation to support feature interpretation.
