In this paper, we describe an approach for learning planning domain models directly from natural language (NL) descriptions of activity sequences. The modelling problem has been identified as a bottleneck for the widespread exploitation of various technologies in Artificial Intelligence, including automated planners. There have been great advances in modelling assisting and model generation tools, including a wide range of domain model acquisition tools. However, for modelling tools, there is the underlying assumption that the user can formulate the problem using some formal language. And even in the case of the domain model acquisition tools, there is still a requirement to specify input plans in an easily machine readable format. Providing this type of input is impractical for many potential users. This motivates us to generate planning domain models directly from NL descriptions, as this would provide an important step in extending the widespread adoption of planning techniques. We start from NL descriptions of actions and use NL analysis to construct structured representations, from which we construct formal representations of the action sequences. The generated action sequences provide the necessary structured input for inducing a PDDL domain, using domain model acquisition technology. In order to capture a concise planning model, we use an estimate of functional similarity, so sentences that describe similar behaviours are represented by the same planning operator. We validate our approach with a user study, where participants are tasked with describing the activities occurring in several videos. Then our system is used to learn planning domain models using the participants' NL input. We demonstrate that our approach is effective at learning models on these tasks.

Domain model acquisition is the problem of learning the structure of a state-transition system from some input data, typically example transition sequences. Recent work has shown that it is possible to learn action costs of PDDL models, given the overall costs of individual plans. In this work we have explored the extension of these ideas to narrative planning where cost can represent a variety of features (e.g. tension or relationship strength) and where exact solutions don’t exist. Hence in this paper we generalise earlier results to show that when an exact solution does not exist, a best-fit costing can be generated. This approach is of particular interest in the context of plan-based narrative generation where the input cost functions are based on subjective input. In order to demonstrate the effectiveness of the approach, we have learnt models of narratives using subjective measures of narrative tension. These were obtained with narratives (presented as video in this case) that were encoded as action traces, and then scored for subjective narrative tension by viewers. This provided a collection of input plan traces for our domain model acquisition system to learn a best-fit model. Using this learnt model we demonstrate how it can be used to generate new narratives that fit different target levels of tension.

Narrative generation represents an application domain for AI planning where plan quality is related to properties such as shape of plan trajectory. In our work we have developed a plan-based approach to narrative generation that uses character relationships as a key determinant in controlling plan shape (relationships are key in genres such as serial dramas and soaps). Our approach is implemented in a demonstration Interactive Narrative, called NetworkING, set in the medical drama genre. The system features a user-friendly mechanism for specifying relationships between virtual characters, via a social network and real-time visualisation of generated narratives on a 3D stage.

Many known planning tasks have inherent constraints concerning the best order in which to achieve the goals. A number of research efforts have been made to detect such constraints and use them for guiding search, in the hope to speed up the planning process. We go beyond the previous approaches by defining ordering constraints not only over the (top level) goals, but also over the sub-goals that will arise during planning. Landmarks are facts that must be true at some point in every valid solution plan. We show how such landmarks can be found, how their inherent ordering constraints can be approximated, and how this information can be used to decompose a given planning task into several smaller sub-tasks. Our methodology is completely domain- and planner-independent. The implementation demonstrates that the approach can yield significant performance improvements in both heuristic forward search and Graphplan-style planning.