To interact with humans, artificial intelligence (AI) systems must understand our social world. Within this world norms play an important role in motivating and guiding agents. However, very few computational theories for learning social norms have been proposed. There also exists a long history of debate on the distinction between what is normal (is) and what is normative (ought). Many have argued that being capable of learning both concepts and recognizing the difference is necessary for all social agents. This paper introduces and demonstrates a computational approach to learning norms from natural language text that accounts for both what is normal and what is normative. It provides a foundation for everyday people to train AI systems about social norms.

A longstanding challenge for machine learning is to learn from complex structured examples in broad, open domains. We believe that domain-independent analogical mapping and constraint propagation can form an effective foundation for such learning. Our experience applying these techniques to Tactical Decision Games led us to develop several strategies that make use of limited domain knowledge to assist in the transfer and adaptation of precedents. Although these additional techniques require some domain-specific knowledge, we believe them to be useful in a broad variety of domains. We have been exploring analogical learning as part of developing interactive companion systems (Forbus and Hinrichs, 2004), software agents that learn over the long term. One important aspect of a companion is that it should learn from experience by accumulating examples. This is a weak form of learning that we expect to augment eventually with facilities for generalization, but it is a critical capability nevertheless.

For more effective collaboration, users and autonomous systems should interact naturally. We propose that sketch-based interaction coupled with qualitative representations and analogy provides a natural interface for users and systems. We introduce comic graphs that capture tasks in terms of the temporal dynamics of the spatial configurations of relevant objects. This paper demonstrates, through a strategy simulation example, how these models could be learned by demonstration, transferred to new situations, and enable explanations.

Designers often use a series of sketches to explain how their design goes through different states or modes to achieve its intended function. Learning how to create such explanations turns out to be a difficult problem for engineering students. An automated "crash test dummy" to let students practice explanations would be desirable. This paper describes how to carry out a core piece of the reasoning needed in such system. We show how an open-domain sketch understanding system can be used to enter many aspects of such explanations, and how qualitative mechanics can be used to check the plausibility of the intended state transitions. The system is evaluated using a corpus of sketches based on designs from an engineering school design and communications course.

Artificial Intelligence 51, pp. 95-143

Proc. AAAI-80, pp. 170-73.