Skip to main content

Karen Spärck Jones


I would like to introduce this distinguished scientist from my personal experience as I was a PhD student of John I. Tait who was a PhD student of Karen Spärck Jones.

When I started my PhD studies I asked John Tait about new references in the field of cross language information retrieval (CLIR), he told me to go to the library and look for a book called Synonymy and Semantic Classification.

I went to the library and found the book, the book was old issued in 1986 for a PhD thesis wrote in 1964 as I remember. I thought that John Tait made a mistake, and at that time I did not know who Karen Sparck Jones is.

I took the book back home and started reading, I couldn't stop reading the book until I fall into sleep. I figured out that this book was written by a distinguished person.

After few days I had a meeting with John Tait and I told him do you remember the book that you advised me to read? He told me yes what about it? I told him it is an excellent book, the best I ever read. He told me you wanted a recent reference and I gave you a very old one but you like it? I told him this reference is not old the author had a sight for many years forward.

Then after this John Tait wrote a book (Charting a New Course: Natural Language Processing and Information Retrieval: Essays in Honour of Karen Spärck Jones) and mentioned this story in the preface.

The research carried out by Karen Sparck Jones and her research students leaded the development of the field of information retrieval and the use of semantic resources to disambiguate query terms.

I will quote here some sentences from her PhD thesis (1964) Synonymy and Semantic Classification:

Requirement for a machine translation dictionary:

“I have been particularly concerned with the problem of finding a general method for constructing a semantic dictionary for machine translation. Such a dictionary must, as we have seen, be both precise and explicit; and it must embody some theory about the way in which words are used” (page 14).

“A machine translation dictionary has to be a dictionary in the ordinary sense: it must give definitions or descriptions of the meanings of words. It must also, however, give some indication of the kind of context in which the words are used, that is must be a 'semantic classification' as well as a dictionary” (page 16).

In relationship of semantic relations to linguistic relations:

“We can put some restriction on the notion of semantic relation, so that the problem of defining relations becomes more manageable. I shall say, therefore, that there is a difference between linguistic relations between words and factual relations between what they stand for, which depends on a distinction between the intra-linguistic use of a word and its extra-linguistic reference” (page 49).

Finally when John Tait asked me if I have in mind a name of an external examiner for PhD viva, that was in early May 2007, I told him anyone other than Karen Spärck Jones, he told me don't worry she passed away last month. a tear dropped from my eye.

I loved this lady while I didn't have the honour to meet her, and I wrote this article in honour of someone that I loved and learned a lot from while I never met.

Mustafa Abusalah, 01-05-2009

For more information about Karen please visit the link here.

Comments

Popular posts from this blog

Coldfusion Facebook Graph API publish to your wall and your friends walls

In this tutorial we will learn by full coldfusion Graph API code example how to publish on your wall and your friends walls. This application uses new oauth authentication method. The code is divided into four files: we will first start with a file called index.cfm: <cfoutput>         <!--- Your FB application IDS --->       <cfset api_key = ""/>     <cfset secret_key = ""/>     <cfset appID = ""/>     <!--- create a connection to the fb graph cfc --->     <cfset graphCFC = createObject("component", "graph").init(#appID#, #api_key#, #secret_key#) />     <!--- If user is authenticated or his access token is set create a cookie --->        <cfif not isdefined("cookie.access_token") and isdefined("url.access_token")>         <cfset cookie.acce...

Implementation of Facebook Graph API in Coldfusion

Facebook has launched a new FB API called Graph that simplifies FB applications development, the new Graph API allows website owners to create Single Sign On (SSO) with Facebook and allow websites owners to be able to import a lot of users information but after their permissions. The code below written in Coldfusion gives an example on how to Create FB login/logout button and then how to retrieve the created cookie, and use it for further development:     <body>   <cfoutput>     <!--- Your FB application IDS --->     <cfset api_key = "XXXXXXXXXXXXX"/>     <cfset secret_key = "XXXXXXXXXXXXXX"/>     <cfset appID = "XXXXXXXXXXXXXX"/>     <!--- Facebook login/logout button --->     <p><fb:login-button perms="email,user_birthday" autologoutlink="true"></fb:login-button></p>     <!--- Facebook login/logo...

The New Facebook Wall Publishing method Stream.publish

As recently facebook stopped supporting the feedstory publishing method to publish news on users wall and application developers from now on can't create feed templates. Facebook released a new methods to do so which are: FB.Connect.streamPublish stream.publish Facebook.streamPublish I figured out those new methods when I was building a new facebook application yesterday to predict who will win the soccer match between Egypt and Algeria tonight. It took me sometime to succeed in publishing the story on the wall, so I will share with you my experience and how I solved it out in Coldfusion. As I build my applications using the work of Andrew Duckett who created the Facebook Rest Client Component , I will start from FacebookFBMLCLient.cfc file which contains the methods that communicates with facebook APIs. As I have used one of the three methods above which sounded the best as I guess, this example will be based on stream.publish method. The first step is to add the following method...