We consider here the problem of building a never-ending language learner; that is, an intelligent computer agent that runs forever and that each day must (1) extract, or read, information from the web to populate a growing structured knowledge base, and (2) learn to perform this task better than on the previous day. In particular, we propose an approach and a set of design principles for such an agent, describe a partial implementation of such a system that has already learned to extract a knowledge base containing over 242,000 beliefs with an estimated precision of 74 % after running for 67 days, and discuss lessons learned from this preliminary attempt to build a never-ending learning agent.

INTRODUCTION How does an autonomous agent that interacts with an environment learn to survive in the environment and make the most out of it? More specifically, how can it develop a set of coping skills that are highly specific (geared toward very particular situations) and thus highly efficient but, at the same time, acquire sufficiently general knowledge that can be readily applied to a variety of different situations? Although humans seem to possess such abilities and seem to be able to achieve an appropriate balance between the two sides, existing AI systems fall far short. There has been a great deal of work demonstrating the difference between procedural knowledge and declarative knowledge (or conceptual and subconceptual knowledge; e.g., Anderson 1982, 1990, Keil 1989, Damasio et al. 1990, Sun 1994). It is believed that a balance of the two is essential to the development of complex cognitive agents. For example, one way to learn a sequential decision task, such as navi

We consider here the problem of building a never-ending language learner; that is, an intelligent computer agent that runs forever and that each day must (1) extract, or read, information from the web to populate a growing structured knowledge base, and (2) learn to perform this task better than on the previous day. In particular, we propose an approach and a set of design principles for such an agent, describe a partial implementation of such a system that has already learned to extract a knowledge base containing over 242,000 beliefs with an estimated precision of 74 % after running for 67 days, and discuss lessons learned from this preliminary attempt to build a never-ending learning agent.