The Ohio State University

TechTips: WolframAlpha Computational Engine

Wolfram Alpha is a ‘computational knowledge engine’ that answers questions, solves equations, and cross-references data types by synthesizing it into different combinations. The goal of WA is to make systematic knowledge immediately computable by drawing on terabytes of curated data.

The service is the brainchild of Stephen Wolfram, who is behind the Mathematica computational software and author of the book  A New Kind Of Science. Mathematica includes data is curated by more than 100 Wolfram employees whom over the years have built a large knowledge base.

Now, you can have access to it as well.

People usually approach and compare WA to Google when they first use it. They type in standard queries, such as their own name, and quickly become frustrated when there are no results. However, WA is NOT a search engine. The service is not trying to out-Google Google.

It doesn’t index Web sites, scans keywords in a search query, and return a list documents that may contain an answer like Google does. It isn’t a community-sourced knowledge base, like Wikipedia. It doesn’t parse natural language and then retrieve documents from other data sources, like Powerset.

Instead, WA reads a question in natural language, scans its databases, and provides an answer using facts, figures, models, charts, and illustrations. It actually computes an answer to a question. WA presents all the results in its own interface rather then sending the searcher onto other Web sites.

The service is available for ad hoc, personal, non-commercial use only. For such use anyone is welcome to download results, print copies, and store downloaded content on your computer.

It’s current knowledge base makes WA is a neat niche service/tool/toy for scientists and technologists. At this early point there are going to be questions about how everyday Web users would benefit from using it over other resources. Still, it’s a very interesting concept which needs input and feedback from those playing with it.

Here are a few search examples:

• Ohio State University
• myoglobin
• where is the international space station?
  
• how far is ann arbor from columbus?

There are all sorts of Easter Eggs as well, such in the response to 88mph and to the question “how many licks does it take to get to the center of a tootsie pop?”

- Eric Schnell

Add comment May 19th, 2009

TechTips: Searching Wikipedia Using Powerset

Powerset is a discovery tool that (currently) extracts and organizes information contained  in Wikipedia and Freebase. A natural language algorithm allows the searcher to use keywords, phrases, or even simple questions.  The results page aggregates information from across multiple articles, in essence creating an enhanced Wikipedia article.For example, here is the search results for Thompson Library.Powerset (purchased by Microsoft in 2008) allows users to highlight passages which persist when the URL is bookmarked or shared the URL.  Once the URL is copied and shared, the modified page will automatically scroll to the highlighted section. It’s great for sharing notes on Twitter or posting a specific sentence in context into a blog.For example, let’s say that you wanted to share that information that it has been reported that the size of the Thompson Library is expected to decrease 20% when it re-opens after the current renovation. As opposed to just sending that bit of information, one can send a link that directs others to the highlighted text. Now the the fact can be read in context of the Wikipedia article.The service also creates ‘Factz’ which are representations of information extracted from sentences. They are represented in three parts: the subject, relation and objects. Factz will appear for general topic queries on the search results page. On a topic page, Factz extracted from a given page in the article outline.The following video provides a nice introduction to the service.-Eric Schnell 

1 comment May 7th, 2009