2025-11-18T16:04:13.800952

FRIREN: Beyond Trajectories -- A Spectral Lens on Time

Wang

Long-term time-series forecasting (LTSF) models are often presented as general-purpose solutions that can be applied across domains, implicitly assuming that all data is pointwise predictable. Using chaotic systems such as Lorenz-63 as a case study, we argue that geometric structure - not pointwise prediction - is the right abstraction for a dynamic-agnostic foundational model. Minimizing the Wasserstein-2 distance (W2), which captures geometric changes, and providing a spectral view of dynamics are essential for long-horizon forecasting. Our model, FRIREN (Flow-inspired Representations via Interpretable Eigen-networks), implements an augmented normalizing-flow block that embeds data into a normally distributed latent representation. It then generates a W2-efficient optimal path that can be decomposed into rotation, scaling, inverse rotation, and translation. This architecture yields locally generated, geometry-preserving predictions that are independent of the underlying dynamics, and a global spectral representation that functions as a finite Koopman operator with a small modification. This enables practitioners to identify which modes grow, decay, or oscillate, both locally and system-wide. FRIREN achieves an MSE of 11.4, MAE of 1.6, and SWD of 0.96 on Lorenz-63 in a 336-in, 336-out, dt=0.01 setting, surpassing TimeMixer (MSE 27.3, MAE 2.8, SWD 2.1). The model maintains effective prediction for 274 out of 336 steps, approximately 2.5 Lyapunov times. On Rossler (96-in, 336-out), FRIREN achieves an MSE of 0.0349, MAE of 0.0953, and SWD of 0.0170, outperforming TimeMixer's MSE of 4.3988, MAE of 0.886, and SWD of 3.2065. FRIREN is also competitive on standard LTSF datasets such as ETT and Weather. By connecting modern generative flows with classical spectral analysis, FRIREN makes long-term forecasting both accurate and interpretable, setting a new benchmark for LTSF model design.

academic

FRIREN/FERN: Beyond Trajectories -- A Spectral Lens on Time

基本信息

论文ID: 2505.17370
标题: Chaining Spectral Pearls: Ellipsoidal Forecasting Beyond Trajectories for Time Series
作者: Qilin Wang (Independent Researcher)
分类: cs.LG
发表时间: October 14, 2025 (arXiv preprint v2)
论文链接: https://arxiv.org/abs/2505.17370

注意: 根据PDF内容，该论文实际名为"FERN (Forecasting with Ellipsoidal RepresentatioN)"，摘要中的"FRIREN"似乎是早期版本名称。

摘要

当前长期时间序列预测(LTSF)实践主要关注随机数据上的逐点指标，掩盖了确定性混沌下的脆弱性。本文提出在经典混沌系统上进行压力测试，并预测未来几何结构而非精确轨迹。FERN是一个几何感知的预测器，采用每个patch的局部线性传输和显式谱因子(特征向量/特征值)，产生结构保持预测和稳定性、模式、政权转换的可操作诊断。除了MSE/MAE，还报告切片Wasserstein距离(形状保真度)和有效预测时间(水平稳定性)。在Lorenz63、Rössler和Chua系统上，FERN相比强LTSF基线提供了显著更低的误差和改进的稳定性，同时在ETT和Weather上保持竞争力。

研究背景与动机

问题定义

核心问题: 现有LTSF模型在确定性混沌系统下表现脆弱，过度关注逐点预测精度而忽视几何结构保持
评估盲点: 标准评估协议存在两个盲点：
- 过度奖励周期性/噪声数据上的模型，忽视混沌下的脆弱性
- 过度强调逐点误差(MSE/MAE)，忽视几何保真度

研究动机

实际需求: 长期预测必然失败，但黑盒模型缺乏诊断失败模式的工具，影响信任和采用
理论基础: 基于Takens嵌入定理，单通道时间延迟嵌入可重构动力系统的拓扑等价吸引子
几何观点: 提出"目标条件局部几何，而非动力学"的新预测哲学

核心贡献

新评估协议:
- 在低维混沌系统上进行压力测试
- 引入几何感知的补充指标(Wasserstein/SWD)
- 提出有效预测时间(EPT)量化可靠预测边界
新预测哲学:
- 目标条件局部几何而非动力学
- 通过椭球链("珍珠串")保持吸引子形状
- 提供几何不确定性表示
FERN模型:
- 融合Normalizing Flows、Optimal Transport和Koopman算子技术
- 实现UΛU⊤ + t形式的Brenier式映射
- 提供完整谱透明度用于失败模式分析

T(y) ≈ T(y₀) + J_T(y₀)(y - y₀) = UΛU⊤y + (T(y₀) - UΛU⊤y₀)

其中：

U: 正交旋转矩阵(特征向量)
Λ: 对角非负缩放矩阵(特征值)
残差项: 平移

2. Koopman增强

U(z)Λ(z)U(z)⊤ → U(z)KΛ(z)K⊤U(z)⊤

其中K是固定可学习的2×2块对角矩阵a -b; b a，模拟复值特征值。

3. 宏观结构：ANF扩展

采用编码器-传输器架构：

算法1: 编码器(X ↔ Z)和椭球传输层

1. z ← N(0,I); y₀ ← N(0,I)
2. for i=1 to K_enc=5:
   - z ← s*(x) ⊙ z + t(x)  # x→z scale-shift
   - x ← s*(z) ⊙ x + t(z)  # z→x scale-shift
3. y_rot ← KU(z)y₀         # 旋转和自旋缩放
4. y_scaled ← Λy_rot       # 非负各向异性缩放
5. y_unrot ← U(z)⊤K⊤y_scaled # 旋转回来
6. y* ← y_unrot + t(z)     # 平移