提出可连续演化的因果机制建模方法，能精准追踪动态系统的变化轨迹。
TRACE: Trajectory Recovery for Continuous Mechanism Evolution in Causal Representation Learning
- 用凸组合建模机制的连续过渡，混合系数随时间变化
- 实验显示轨迹恢复相关性高达0.99，显著优于离散切换模型
- 适合研究连续演化系统的因果推理，如车辆运动或步态变化
时间因果表示学习方法通常假设因果机制在离散域间瞬时切换，但现实系统常呈现连续机制演变。例如车辆转弯时动力学渐变，人从走路到跑步的步态平滑过渡。本文将此设定形式化为有限个原子机制的凸组合，由时变混合系数控制。理论证明潜变量与连续混合轨迹可联合识别。进一步提出TRACE框架——基于专家混合模型，每个专家学习一个原子机制，测试时可恢复机制演化轨迹。该方法能泛化至训练中未见过的中间状态。在合成与真实数据上的实验表明，TRACE恢复的混合轨迹相关性最高达0.99，显著优于离散切换基线。
原文摘要 · Abstract (English)
Temporal causal representation learning methods assume that causal mechanisms switch instantaneously between discrete domains, yet real-world systems often exhibit continuous mechanism transitions. For example, a vehicle's dynamics evolve gradually through a turning maneuver, and human gait shifts smoothly from walking to running. We formalize this setting by modeling transitional mechanisms as convex combinations of finitely many atomic mechanisms, governed by time-varying mixing coefficients. Our theoretical contributions establish that both the latent causal variables and the continuous mixing trajectory are jointly identifiable. We further propose TRACE, a Mixture-of-Experts framework where each expert learns one atomic mechanism during training, enabling recovery of mechanism trajectories at test time. This formulation generalizes to intermediate mechanism states never observed during training. Experiments on synthetic and real-world data demonstrate that TRACE recovers mixing trajectories with up to 0.99 correlation, substantially outperforming discrete-switching baselines.
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。