January 2, 2007

hakia, a Knowledge Search Engine

hakia is a new search engine that tries to understand more from the web pages and focuses on information rather than web pages. "The basic promise is to bring search results by meaning match - similar to the human brain's cognitive skills - rather than by the mere occurrence (or popularity) of search terms."

For simple queries like [panda] or [Napoleon Bonaparte], hakia shows galleries: a big list of results grouped by topics that try to answer to many possible queries. For Napoleon Bonaparte, you'll see these clusters: official websites, biographies, pictures, news and interviews, fan sites, awards, speeches, myths, controversies, resources, innovations, statistics, bibliography, filmography. So it's like an automatically-generated Wikipedia page, without a coherent story.

Like other search engines, hakia shows snippets for each result, but these snippets are bigger and try to highlight a possible answer to your query. Unfortunately, hakia is not always on the right track and when it fails, it fails completely.

hakia works only for English queries, doesn't recognize spelling errors too often, doesn't have advanced operators and it's a bad choice for navigational queries (the first search result for [google] is google.org).

The search engine will get out of beta this year and its founders predict it will be the first engine to use ontological semantics that enable it "to perceive concepts beyond words and retrieve results with meaningful equivalents".

2 comments:

  1. This seems similar to VIVISIMO, which clusters or categrorizes searches on the fly:
    http://search.vivisimo.com/

    ReplyDelete
  2. Not quite, they show clusters only for general queries and, unlike Vivisimo, this is not a meta-search engine. It still needs a lot of work, because for some queries they do a better job than Google and other search engines, but for most of them, the results are worse.

    ReplyDelete

Note: Only a member of this blog may post a comment.