Analysing input in dialog systems

Abstract

One of the main limitations in existent domain-independent conversational agents is that the general and linguistic knowledge of these agents is limited to what the agents' developers explicitly defined. Therefore, a system which analyses user input at a deeper level of abstraction which backs its knowledge with common sense information will essentially result, in a system that is capable of providing more adequate responses which in turn result in a better overall user experience. From this premise, the aims of this project were: • To build an expandable, customisable, flexible and modular proof-of-concept system which provides a syntactic, semantic and pragmatic understanding of input. • To simulate intelligence by providing adequate output and logical conclusions derived from dialogue input and local and external knowledge bases. A prototype was implemented upon the proposed framework, which makes use of various natural language processing tools, online and offline knowledge bases, and other information sources, to enable it to comprehend and construct relevant responses aided by these tools. The resulting system was evaluated by the use of an implemented prototype with a specified set of boundaries for which people rated and commented on conversation transcripts carried out with the system. It was concluded that the system performed favorably when provided with input it was designed to comprehend, and for which it had access to information from the available information sources, but still requires further refinement as to be able to process more linguistically and logically complex input.