Scooter: A Personal Knowledge Retrieval Agent For The Web.
There has been a proliferation of search engines feeding the insatiable appetite of the users of the World Wide Web for data.
After a decade of use, it was decided to add semantics to the Web, thereby transforming data into information; and so the "Semantic Web" was born. Unfortunately at this juncture in its life, the Semantic Web has degenerated into a set of standards, markup languages and related processing tools; sidelining the promise of semantics.
The "Scooter" (inspired by the Latin word scitor, meaning to seek, to know, inquire, ask) project takes the Web to the next logical step, the "Pragmatic Web"; turning information into knowledge by harnessing Artificial Intelligence techniques, Expert System ,Data Mining and knowledge Sharing technologies.
The ubiquity and pervasiveness of the cellphone (in all its variant forms) makes it ideal as one's "Personal Knowledge Retrieval Agent".
The marriage of these hardware and software technologies will provide "just in time" knowledge delivery through a low cost, easy-to-use interface, either as a subscriber or pay-per-use service.
Instead of dumping endless "page hits" to users in hopes that they can find "the wheat amongst the chaff" or forcing sponsor "ranked" pages on the user, the Scooter system searches its Knowledge Bases and responds to user natural language queries with natural language responses. For time and location sensitive queries, the Scooter system has engines that can "glean" the required knowledge from the Web. If all else fails, Scooter will delegate the query to a human expert.
Rather than forage for data on the Web, the Scooter system harvests information from predetermined sites. This information is then ontologically mapped into Knowledge Bases.
In order to make the knowledge retrieval process as simple and "user friendly" as possible, the front end of the Scooter system is an amanuensis your personal intelligent agent. A user converses with the intelligent agent through natural language, either text or speech. The agent knows the best modality to present responses, either text, voice, image, video etc., based on such criteria as the format of the response delivery vehicle (cellphone, PDA, PC, etc.), cost, bandwidth, and so forth.
For devices capable of permanently housing the intelligent agent, an avatar is used. This avatar is capable of becoming a "friend". It can learn a user's likes, dislikes, preferences, fetishes, etc. and it will tailor future conversations to match.
The prototypical system allows a user to text message a question to the "Scooter" system, in the domain of professional hockey, and receive a text message response.