812
个编辑
更改
→粗粒化与时间演化算子的可交换性
<ref name=":24">Otsuka, Jun, and Hayato Saigo. "Process theory of causality: a category-theoretic perspective." Behaviormetrika 51.1 (2024): 21-36.</ref>
<ref name=":24">Otsuka, Jun, and Hayato Saigo. "Process theory of causality: a category-theoretic perspective." Behaviormetrika 51.1 (2024): 21-36.</ref>
<ref name=":25">Zennaro, F. M., Turrini, P., & Damoulas, T. (2023). Quantifying consistency and information loss for causal abstraction learning. arXiv preprint arXiv:2305.04357.</ref>
<ref name=":25">Zennaro, F. M., Turrini, P., & Damoulas, T. (2023). Quantifying consistency and information loss for causal abstraction learning. arXiv preprint arXiv:2305.04357.</ref>
<ref name=":25">Beckers, S., Eberhardt, F., & Halpern, J. Y. (2020, August). Approximate causal abstractions. In Uncertainty in artificial intelligence (pp. 606-615). PMLR.</ref>
<ref name="ApproxCA">Beckers, S., Eberhardt, F., & Halpern, J. Y. (2020, August). Approximate causal abstractions. In Uncertainty in artificial intelligence (pp. 606-615). PMLR.</ref>
。满足一致性的粗粒化方案能够充分保留动力学中的有效信息，并丢弃无用信息。具体来说，宏观状态已经捕捉了与未来演化相关的所有“慢变量”或“相关信息”，而被粗粒化掉的那些微观细节，只影响瞬时噪声或快速弛豫过程，不会长期、系统性地改变宏观演化。
。满足一致性的粗粒化方案能够充分保留动力学中的有效信息，并丢弃无用信息。具体来说，宏观状态已经捕捉了与未来演化相关的所有“慢变量”或“相关信息”，而被粗粒化掉的那些微观细节，只影响瞬时噪声或快速弛豫过程，不会长期、系统性地改变宏观演化。