Answering Hindsight Queries with Lifted Dynamic Junction Trees

TL;DR

This paper extends the lifted dynamic junction tree algorithm (LDJT) to efficiently handle smoothing and hindsight queries in probabilistic relational temporal models, enabling faster multiple temporal query answering.

Contribution

We introduce an extension to LDJT for smoothing, including an efficient backward pass and various instantiation options, making hindsight queries feasible from the start.

Findings

LDJT with the new backward pass improves smoothing efficiency.

The relational forward-backward algorithm enables hindsight queries from the initial time.

LDJT outperforms static algorithms in multiple temporal query scenarios.

Abstract

The lifted dynamic junction tree algorithm (LDJT) efficiently answers filtering and prediction queries for probabilistic relational temporal models by building and then reusing a first-order cluster representation of a knowledge base for multiple queries and time steps. We extend LDJT to (i) solve the smoothing inference problem to answer hindsight queries by introducing an efficient backward pass and (ii) discuss different options to instantiate a first-order cluster representation during a backward pass. Further, our relational forward backward algorithm makes hindsight queries to the very beginning feasible. LDJT answers multiple temporal queries faster than the static lifted junction tree algorithm on an unrolled model, which performs smoothing during message passing.