Powerset launches today to an overzealous amount of fanfare, as the mysterious new semantic search technology will open its doors for the first time to the public. The public must understand however, that Powerset is not a finished product, as search and technology is an evolving discipline. Consider today’s launch Powerset’s first Powerstep.
Over the past year or so, the words “Powerset” and “Google Killer” have been breathed in the same breath by way too many in the blogosphere and news corporations. So much in fact, the company has tried to step back from the nearly impossible expectation as being the next Google, and concentrate on its core offerings, a new dimension in search and navigation. Powerset’s search and discovery functionality is different and does lend quite an efficient measure to searching; Powerset will make a difference.
Natural Language Search
Powerset has launched their core semantic research engine today which indexes and searches Wikipedia and Freebase. Powerset claims to reinvent how users search and discover information, and they are taking the right steps in the direction to meet that goal.
Powerset’s semantic technology is a hybrid of patents licensed from PARC (formerly Xerox PARC) and Powerset’s proprietary research; and unlike traditional search technologies from Google, Yahoo and Teoma; which look for words, Powerset matches the meaning of the users’ query to the meaning of sentences.
Natural language search attempts to serve more relevant results for user queries which go beyond the basic keyword model, serving results with higher precision and better recall, due to the searcher actually understanding the matching of results to the query.
Rediscovering Wikipedia with Powerset
Today, Powerset unveils the first public usage of their search technology, by giving its users the ability to search over 2.5 million Wikipedia articles. Yes, this is a lot of fanfare for a search technology which searches Wikipedia, or gives Wikipedia a new dimension of discovery.
“Wikipedia serves a fundamental human need – a single destination to find a broad range of information,” according to Barney Pell, Powerset co-founder and CTO. “This broad content base is a great way for us to showcase Powerset’s functionality while giving people a faster time to satisfaction.”
But what exactly is Powerset’s functionality?
From my experience Powerset associates different terms, nouns, verbs, subjects and their synonyms with search results and groups occurances of those terms within Wikipedia. For example, a search for “politicians who dies of disease” will result in indexed pages of Wikipedia with occurrences of not just the keyterm ‘politicians’ but the actual name of politicians. Then, the technology will also associate terms like killed, and then associate disease with different diseases themselves, like cancer or Alzheimers.
Then, Powerset delivers a list of those politicians in its results, and the diseases they died from. Google, on the other hand, matches the exact terms within its index and delivers results only of instances of those terms, not the true meaning behind the searcher intent. So, it could be said that Powerset delivers answers (hey, wasn’t this the original idea behind Ask Jeeves?) while Google deliver results.
One Tiny Step for Powerset, One Giant Leap for the Semantic Web
Powerset is more or less dependent upon open source third party authorities like Wikipedia (under GNU Free Documentation) and Freebase (under Creative Commons) for its current informational index of content, as unlike a common search engine which delivers users to external sites; for now Powerset is keeping them to the confines of their informational partners. Over time, more and more of these partners will probably be added to the Powerset experience as once again, this first Powerstep is but one of many babysteps for the company.
As a research tool, I find Powerset’s searching of Wikipedia beneficial for searching for events in the past (unless they are ‘Wikiality’) but for searching current events, Powerset says that there is a time lapse between when content is uploaded live to Wikipedia and when Powerset updates its ‘index’. What I would like to see here is the ability for users to upload their OPML files into the Powerset system or have some kind of Bloglines plugin so I can perform research on all of the blog posts in an archive of blogs and news sources I trust.
I may also be jumping the gun here but I could also see Powerset technology as a natural compliment to Yahoo! channels, and the content Yahoo syndicates within its channels from inside and outside of Yahoo, but all within Yahoo. Google is dependent upon the keyword search model and has issues with the incompetence of semantic search, Yahoo on the otherhand has dramatically opened up its site to outside developers, somewhat recognizing semantic with its use of Hadoop, which Powerset also supports. Not to start any rumors, but the usage of Powerset’s semantic researching with Yahoo would be a nice partnership.
Besides Powerset, Hakia is more or less the leader in Semantic search, but it takes a different approach in building its own proprietary index instead of depending on a trusted third party source, as Powerset does with Wikipedia or Freebase. (Here’s a nice comparison of the two technologies by Phil Butler at ReadWriteWeb)
If you’d like to learn more about Semantic Search, here are some other major Semantic Search Engines :
Just the Factz
Getting back to Powerset, besides their semantic natural language approach, Powerset is also categorizing their search results in five different blended and branded areas :
- Factz – When users enter a topic query, Powerset assembles a compact summary of interesting, and sometimes surprising Factz, extracted from pages across Wikipedia
- Dossiers – Powerset creates a summary of information found in Freebase and Wikipedia to give users a quick overview about a topic
- Answers – For many questions, Powerset automatically assembles an answer list from sentences in Wikipedia or data in Freebase
- Semantic Highlighting – The most relevant search results are highlighted based on the meaning of a users’ question.
- Minibrowser – A result can be expanded to show the snippet in the context of the full Wikipedia article
The future of Powerset is up in the air, but after today the online world will have a better grasp of what Barney Pell hope to achieve, and hopefully more awareness of semantic search.