Online Learning of Event Definitions

TL;DR

This paper introduces an online ILP system for learning Event Calculus theories in real-time, enabling efficient event recognition with competitive accuracy and faster training compared to batch methods.

Contribution

It presents a novel online learning approach for Event Calculus theories using Hoeffding bounds and abductive ILP, allowing single-pass, real-time event recognition.

Findings

Achieves comparable predictive accuracy to batch methods.

Significantly faster training times in online setting.

Outperforms hand-crafted rules in activity recognition.

Abstract

Systems for symbolic event recognition infer occurrences of events in time using a set of event definitions in the form of first-order rules. The Event Calculus is a temporal logic that has been used as a basis in event recognition applications, providing among others, direct connections to machine learning, via Inductive Logic Programming (ILP). We present an ILP system for online learning of Event Calculus theories. To allow for a single-pass learning strategy, we use the Hoeffding bound for evaluating clauses on a subset of the input stream. We employ a decoupling scheme of the Event Calculus axioms during the learning process, that allows to learn each clause in isolation. Moreover, we use abductive-inductive logic programming techniques to handle unobserved target predicates. We evaluate our approach on an activity recognition application and compare it to a number of batch learning techniques. We obtain results of comparable predicative accuracy with significant speed-ups in training time. We also outperform hand-crafted rules and match the performance of a sound incremental learner that can only operate on noise-free datasets. This paper is under consideration for acceptance in TPLP.