Perennial, Permuted, and Pervasive Search in Ambient Intelligence

Searching in ambient intelligence is complex be- cause it requires fine-grained tracking of human intentions, adapting to varying interests and intentions, and maintaining search accuracy in the context of fast content changes and increasing diversity of information. The paper describes details on some of the computational challenges by framing them at three levels: 1) a human interacting with a machine, 2) a small group of humans and machines collaborating, and 3) decision- support based on diverse and fast evolving information in the context of a large community of users and machines. Along with the challenges, some early glimpses of potential solutions are also discussed.