Vision

The vision of the Comirit project is to pragmatically create useful systems that have a deep understanding of the world. Our objective is not to create replicas of human intelligence, but to create useful tools (in whatever form they may take) for software and robot engineering.

Applications First

Our process is driven primarily by applications of the technology. We believe that focus on real applications has two important benefits:

• Successful applications drive interest in the technology that will help support future research.
• Concrete applications provide a robust yardstick and practical grounding to ensure the research program remains on track. That is, evolving increasingly capable applications is a potentially slower, but, we believe, less risky approach than one that attempts to build powerful artificial general intelligences from scratch.

Intelligent Middleware

In enterprise development, 'middleware' is seen as the 'glue' that managers the data and communications across an enterprise architecture. Our vision is for Comirit to be deployable as an 'intelligent middleware'. Software agents, software applications and robotic systems will be built on top of Comirit: the intelligent middleware will be invoked to solve complex problems, and the middleware itself may perform work as it deems appropriate to react to a complex and changing environment.

Community

We see Comirit as part of a worldwide community of resurging interest in automated commonsense reasoning and artificial general intelligence. While our long term vision remains ambitious, we draw on lessons learnt over the past 50 years of AI research and recognize that many incredibly difficult challenges remain: progress will be necessarily slow and our expectations and claims should be tempered with caution.

The following links are a starting point for further information about the history, future and vision of commonsense reasoning.

• AAAI Commonsense Reasoning: Good introduction to commonsense reasoning.
• Programs with Common Sense: John McCarthy's pioneering paper on building commonsense-aware systems, that remains significant to this day.
• Artificial General Intelligence Research Institute: Community resource for AGI researchers.
• Commonsense Problem Page: Benchmark problems.
• Cyc: Large scale commonsense knowledge engineering.
• Project Halo: Domain specific knowledge engineering.
• Open Mind Commonsense: User-contributed "open" commonsense.
• AGI-08: Upcoming conference on artificial general intelligence.
• Commonsense 2007: Conference on logical commonsense reasoning.

Authored by Benjamin Johnston for the Comirit Project, 2010.