Information Retrieval Radar – Ping 1

It technology terms Information Retrieval (IR) is still at the “toddler” stage of development.  However, rapid progress is starting to be made and there are many innovative and exciting IR techniques and tools emerging to help us tackle the information overload.  These new approaches are not just to be found in the large, complex corporate application suites but also in a plethora of small and innovative companies and increasingly in open source environments.

For individuals and organisation part of the challenge is know to and understand “what is out there”, and that is in itself an IR challenge.  At i-logue we monitor the technologies that enable us to help our clients manage information and therefore we like to maintain a “radar picture” for IR developments.  This is the first post and in a series where I will identify and describe some the techniques and tools we have “spotted on the radar”.  Increasingly IR capabilities can be provided without the need for complex and lengthy development programmes and their significant cost.  In the post I will, where possible, provide examples and links to illustrate techniques and tools.  Inevitably, this will place a bias towards IR in relation to the Internet and the World Wide Web.  The enterprise IR solutions generally have the same capabilities but their integration into enterprise products tends to lag behind what is happening in the smaller, innovative more dynamic IR companies.

So, here are some interesting capabilities that we have spotted on the radar.  Some work on the web and some can be downloaded.  All of them will help you understand a little bit more about IR.

Apture

This really clever browser extension (IE, Firefox, Chrome and Safari) which provides context related in-line search results as you read an article on a web site.  As you read an article on any website you can highlight any word and Apture will conduct a search and return a pop-up window alongside the highlighted word with information drawn from the web major search engines, social feeds, video and images.  Words that have been regularly searched by others are already highlighted discreetly and can be examined with a single click.

I have been using Apture for a couple of weeks now and it really works well.  It significantly speeds up any research task and even makes browsing a more informed experience.   It uses Wikipedia well and also the Google and Bing search engines. The Twitter search is reasonable and is probably based on an individual having a proper Twitter name rather than a pseudonym.  It is not clear when it uses one search engine rather than the other but the results are generally relevant.  Having video and images right there on tabs is also a great productivity gain.  More detailed information, if available, is presented through horizontally scrolling windows.

In addition to the browser plug-ins you can apply Apture to your own web site.  This site is Apture enabled and you should also take a look at The Scientific American’s website.

Infomous 

We are all familiar with “tag clouds” by now.  Those little jumbles of words beloved by news sites and blogs. The size of the words typically relates to either how often a document has been “tagged” with the word or they represent the frequency of a word as a search term on the site.  Infomous looks like a graphical tag cloud but it is much more useful as an IR support tool. Here is how they describe it:  “Infomous summarizes the content of one or more documents and displays the results as an interactive cloud of words of varying sizes and connections.  Each word on the screen is a topic that stands out because it is prominently discussed in the documents or news feeds you are reading. The size of each word reflects the frequency with which it appears in the source, while links between words show proximity in the source.”

Selecting a node from the cloud generates a drop-down showing all the source articles as clickable links.  You can generate Infomous clouds from a website’s content, from a number of external links such a blogs or even Twitter names or hashtags.  The great thing about the approach is that you can very quickly establish trends and relationships between topics. You can customise and embed Infomous clouds in your own website or build personal clouds for research.  A nice public example is the cloud used on The Economist newspaper’s Babbage technology blog.

Hopefully you will have been inspired by these two examples.  Information Retrieval is not just about a keyword and a search box.  There are many more exciting capabilities and we will keep this post as a regular feature up update readers on what we think is innovative and potentially game changing in the IR world.