i-Seek : an intelligent system for eliciting and explaining knowledge

Abstract

We propose i-Seek, an Intelligent System for Eliciting and Explaining Knowledge that leverages the OpenMind [1] Commonsense knowledge base in conjunction with domain- specific knowledge in Personal Finance, Technical Help, and Health domains to act as an advisory system for novice users. Most of the interfaces are plagued by recurrent key problems: 1) elicitation - how to ask questions that enable the expert model to make decisions, and at the same time, are understandable to the novice, and 2) explanation - how to explain rationale behind expert decisions in terms that the user can understand. i- Seek maps the user's goals and expectations to the corresponding expert model's attributes as expressed in domain-specific terms. For example, instead of asking "What is your risk tolerance?", where the user might not comprehend the notion of risk tolerance, i-Seek tries to elicit the same information by asking a non-direct question such as "Do you usually buy lots of lottery tickets?". i-Seek constructs the novice user model by taking into account the user's personal information, interactions history, and the current context.