A natural language interface for a 2D networked game

Authors:
Andrea Corradini;Adrian Bak;Thomas Hanneforth
Affiliations:
University of Potsdam, Computational Linguistics Department, Potsdam, Germany;TV2, Interaktiv, Odense, Denmark;University of Potsdam, Computational Linguistics Department, Potsdam, Germany
Venue:
HCI'07 Proceedings of the 12th international conference on Human-computer interaction: applications and services
Year:
2007

Citing 10
Cited 0

Heuristic evaluation of user interfaces

CHI '90 Proceedings of the SIGCHI Conference on Human Factors in Computing Systems
Heuristic evaluation

Usability inspection methods
A program to solve the Pentomino problem by the recursive use of macros

Communications of the ACM
The Open Agent Architecture

Autonomous Agents and Multi-Agent Systems
Practical experiments with regular approximation of context-free languages

Computational Linguistics - Special issue on finite-state methods in NLP
Unification-based multimodal integration

ACL '98 Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics and Eighth Conference of the European Chapter of the Association for Computational Linguistics
Put My Galakmid Coin into the Dispenser and Kick It: Computational Linguistics and Theorem Proving in a Computer Game

Journal of Logic, Language and Information
CLAWS4: the tagging of the British National Corpus

COLING '94 Proceedings of the 15th conference on Computational linguistics - Volume 1
Natural language and inference in a computer game

COLING '02 Proceedings of the 19th international conference on Computational linguistics - Volume 1
Synthetic Worlds: The Business and Culture of Online Games

Synthetic Worlds: The Business and Culture of Online Games

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes a system for playing a digital version of a board game using natural language. We consider the control of the game through players interacting with a 2D graphical interface by typing in appropriate text instructions. Such an input is syntactically parsed using a set of pushdown transducers to generate a simplified string representation of the original user's sentence. Further, a semantic parser splits the new input representation into a series of frames, each one representing the semantics of the underlying text chunks. Based on the game logic and according to context information, we re-solve ambiguities and incompleteness within these data structures and generate a set of possible game instructions. Eventually, we check preconditions related to the validity of the hypothesized commands and, if the preconditions are met, the equivalent instructions are carried out and the game state is updated. For testing purposes, we fed our system with game instructions that are automatically generated by an equivalent context free grammar that can be described by our pushdown transducers. In the case of parseable instructions, the system has been shown to properly perform with over 98% accuracy. The application can be played over the network based on a client/server architecture where players exchange periodic updates through a central host.