While it is obvious that Search 1.0 was about the user mining into results returned by search engines, the need for moving to a higher plane in the search domain is obvious.
AnswerBus, the answering engine that provides precise and concise answers to queries not in just English but German, French, Spanish, Italian and Portuguese also, is a glimpse into what will be the glue for the web of the future, essentially Search 2.0 powering the semantic web.
Here’s a low down on the working of the AnswerBus engine :
The input for the engine is a simple question from the user that the engine uses to prepare separate queries and mine the popular search engines like Google, Yahoo and MSN Live Search for optimum results. The queries would have to be framed with selective deletions and modifications to the words from the original question.
From the top rated results, the documents are parsed for sentences that match the initial question. The parsing is done by correlating the sentences to HTML tags. After the sentences are obtained from the documents, they are ranked on how well they match to the words in the question asked. Apart from the match-up score, several other procedures are also used (QA specific dictionary, named entities extraction, coreference resolution, and redundancy deletion).
AnswerBus may appear to be a simple information retrieval system. However, there are several characteristics to the working of the engine that I consider are relevant for ushering in the next era of Search.
First, AnswerBus uses the repositories of other search engines which is the equivalent of ‘standing on the shoulders of giants’. This is essential, since Search 2.0 will be more about how well the repositories built by Search 1.0 systems are utilized and intelligently processed for information extraction.
It’s common sense that present day search engines use delegated thinking, the results returned are what other people may have found useful ( they linked to it, not necessarily meaning that they knew exactly what the content was).
With software fast running virtually off the web and semantic web ready to move into wider usage, search 2.0 will be the glue that will bind the applications and let them be discovered on the semantic web. It will have to usher in an era where search engines will return exact queries in various formats which will feed into further processes. Essentially it is about making search results amenable for machine processing, and AnswerBus’s approach marks a beginning, if only a slight one, in that direction.