Mixed-Initiative Assumption-Based Reasoning for Complex Decision-Making

摘要

This paper discusses several critical capabilities of the Disciple-LTA system for complex problem-solving and decision-making, including a transparent and easy to understand reasoning process, a flexible and natural collaboration with the user, and the use of what-if scenarios to cope with incomplete and uncertain information. They allow the user to act as the orchestrator of the reasoning process, guiding the high-level exploration of the decision-making space, while the system implements this guidance by taking into account the user's preferred problem solving strategies, assumptions and biases. These capabilities are discussed in the context of intelligence analysis.