The power of our practical AI agent comes from two central components: A reasoning engine called AICoRE, and a human-like memory called Memory Graph. AICoRE (Adaptive Interactive Cognitive Reasoning Engine) is a cognitive reasoning engine which unifies problem solving and learning. AICoRe fully automates the reasoning process from beginning to end. As such, AICoRE is an Incremental Holistic Human-like Reasoner, covering the full spectrum of reasoning from sensing, reasoning, discovering, planning, learning, remembering until responding and performing.
As the central part of problem solving mechanism, our planner is the most advanced multi-agent cognitive planner dealing with real-world problems among multiple agents.
Unlike most of the classical planners that deal with physical actions only, CMAP is capable of dealing with both physical and cognitive (speech) actions. That is why we refer to our proprietary planner as Cognitive Multi-Agent Planner or CMAP for short. It is implemented in a major planning language called PDDL (Planning Domain Definition Language).
Unlike other PDDL planners, CMAP is highly practical dealing with real world problems while interacting with multiple agents and humans. In addition, it is also integrated with an incremental case-based reasoner over time.