Familiarity Networks and Interactive Television

摘要

Familiarity Networks (fnets) respond to levels of familiarity. Like living systems, fnets are exposed to a continuous stream of input patterns. When an input pattern is similar to patterns seen in the recent past, an fnet responds strongly. When an input pattern is unfamiliar, an fnet responds weakly. This paper describes an application of fnet technology to the improvement of Interactive Television (ITV). After some exposure to the viewer's program preferences, the ITV fnet learns to make good suggestions for upcoming programs. The system moves the viewer quickly to the most probable program choices.