Timeline for Lipschitzness of conditional law of a stochastic filtering problem wrt the Wasserstein distance
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 22, 2022 at 18:10 | comment | added | Justin_other_PhD | @Ilya This is a good point! I have modified the question to focus on the case I'm most interested in, which is conditioning on $\sigma(Y_t)$ not $\sigma(\{Y_s\}_{0\le s\le t})$. Thanks for pointing this out. | |
| Dec 22, 2022 at 18:09 | history | edited | Justin_other_PhD | CC BY-SA 4.0 |
added 130 characters in body
|
| Dec 22, 2022 at 17:42 | comment | added | SBF | I am not an expert on continuous-time Markov processes, but I guess in discrete time if I have observations of $Y$ (being constructed dynamically using $Y$ and $X$), it is not enough for me to just know the latest value of $Y_t$ to get everything I can for the distribution of $X_t$, in most cases knowing the whole trajectory of $Y$ provides a strictly finer conditioning. I guess the same woule apply to continous-time case. Namely, $\Bbb P(X_t|\mathcal F_t^Y)$ is not a function of simply $Y_t$ and $t$. | |
| S Dec 22, 2022 at 16:17 | review | First questions | |||
| Dec 22, 2022 at 16:49 | |||||
| S Dec 22, 2022 at 16:17 | history | asked | Justin_other_PhD | CC BY-SA 4.0 |