|
Abstract : |
Dynamic probabilistic networks (DPNs) are a useful tool for modeling complex stochastic processes. The simplest inference task in DPNs is monitoring (aka filtering) task--- that is, computing a posterior distribution for the state variables at each time step given all observations up to that time. Recursive, constantspace algorithms are well-known for monitoring in DPNs and other models. This paper is concerned with hindsight (aka smoothing)--- that is, computing a posterior distribution given both past and future observations. Hindsight is an essential subtask of learning DPN models from data. Existing algorithms for hindsight in DPNs use O(S, |
