Abstract
Plan recognition deals with reasoning about the goals and execution process of an actor, given observations of its actions. It is one of the fundamental problems of AI, applicable to many domains, from user interfaces to cyber-security. Despite the prevalence of these approaches, they lack a standard representation, and have not been compared using a common testbed. This paper provides a first step towards bridging this gap by providing a standard plan library representation that can be used by hierarchical, discrete-space plan recognition and evaluation criteria to consider when comparing plan recognition algorithms. This representation is comprehensive enough to describe a variety of known plan recognition problems and can be easily used by existing algorithms in this class. We use this common representation to thoroughly compare two known approaches, represented by two algorithms, SBR and Probabilistic Hostile Agent Task Tracker (PHATT). We provide meaningful insights about the differences and abilities of these algorithms, and evaluate these insights both theoretically and empirically. We show a tradeoff between expressiveness and efficiency: SBR is usually superior to PHATT in terms of computation time and space, but at the expense of functionality and representational compactness. We also show how different properties of the plan library affect the complexity of the recognition process, regardless of the concrete algorithm used. Lastly, we show how these insights can be used to form a new algorithm that outperforms existing approaches both in terms of expressiveness and efficiency.
Highlights
A plan recognition algorithm allows an observer to reason about the goals and execution process of an agent, the actor, given a set of its observed actions
Other notable works are Engine for LEXicalized Intent Recognition [ELEXIR, (Geib, 2009)] and YR (Maraist, 2017) which present new approaches for plan library (PL)-based plan recognition algorithms. These algorithms use PL representations that differ both from Probabilistic Hostile Agent Task Tracker (PHATT)’s and from SBR’s. Another possible standard representation that was considered in previous work is the Hierarchical Task Network (HTN) representation used by the Simple Hierarchical Ordered Planner 2 (SHOP2) (Nau et al, 2003)
The basic level actions in the plan library for solving the ROSA problem are the actions that students can perform in the software, such as add new sampler (NS), Create device (CCD), which is executed by adding a device to sampler (SAD), set number of draws in the sampler (SDS) and number of repetitions (SR); SRP, creating a sampler object (CSM) and plotting action (PO) are some of the complex actions
Summary
A plan recognition algorithm allows an observer to reason about the goals and execution process of an agent, the actor, given a set of its observed actions. We create a joint interface that facilitates two representative plan recognition algorithms from the literature, and use this representation in order to provide a thorough, both theoretical and empirical, comparison between them. This comparison enables us to leverage insights from these works and provide a new, improved algorithm that outperforms both baseline algorithms. We will use two plan recognition algorithms from these two families: the graph-based SBR (Avrahami-Zilberbrand and Kaminka, 2005) and the grammar-based PHATT (Geib and Goldman, 2009).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
