1: \begin{abstract}

2: % We study conformal inference with serially dependent data. Our analysis exploits the connection between conformal prediction and permutation tests. To account for the serial dependence in the data, we study block permutation schemes. When the data are exchangeable our conformal inference method enjoys model-free and exact finite sample validity. When exchangeability fails, the proposed approach is approximately valid under weak assumptions on conformity score. 

3: 

4: We extend conformal inference to general settings that allow for time series data. Our proposal is developed as a randomization method and  accounts for potential serial dependence by including  block structures in the permutation scheme such that the latter forms a group. As a result, the proposed method retains the exact, model-free validity when the data are i.i.d.\ or more generally exchangeable, similar to usual conformal inference methods.  When exchangeability fails, as is the case for common time series data, the proposed approach is approximately valid under weak assumptions on the conformity score. 

5: 

6: 

7: 

8: % our approach is robust in that theoretical bounds can be obtained   approximate

9: 

10: \end{abstract}

11: