Archive

Archive for the ‘Computer Industry’ Category

King’s College hosts Semantic Web Event

February 26, 2014 Leave a comment

This is a guest post by Dr Jarred McGinnis, Visiting Research Fellow in the Agents and Intelligent Systems Group of the Department of Informatics at KCL.

It could be my nationality.  I’m American and we’re a chatty lot.  Or, my career. I started out as an academic, a logician and theoretician. Then I moved toward more applied and managerial roles in industry. I’ve spoken to a lot of people with differing perspectives. I’ve come to appreciate the need for dialogue and understanding.

Besides my role as a Visiting Research Fellow here at King’s, lately I’ve been working with Ontotext who provides semantic repositories and semantic NLP solutions. They’re interesting to work with because they are building technologies, often seen as academic and esoteric, to be used in an enterprise setting within governments, organisations and businesses. Often, I am sitting between the technologists and the business owners translating the needs and desires of one party in the terms of the other as they try to decide whether semantics is something their organisation needs.

It was in the spirit of fostering that cross-discipline communication that I put together the Semantic Web Meet Up hosted in King’s incredible Anatomy Lecture Theatre. The title of the discussion was ‘What Linked Data Does, What Linked Data Need’. I wanted students, academics, technologists and business people to be in the same room talking about the same thing. The plan was to gather a number of professionals currently using semantics and discuss what further work is needed to bring this emerging technology to the mainstream. It would have been hard to have a better panel to cover the spectrum of linked data users. We had:

– Fabio Colasanti, Data and Information Architect at EuroMoney, a publisher of highly valuable economics and financial information.
– Tom Heath, Head of Research at ODI, a nonprofit (and technology neutral) organisation encouraging the opening of data, Linked or otherwise.
– Sofia Angeletou, Senior Data Architect at BBC, the publically funded broadcaster we know and love that has spearheaded the use of Linked Data.
– Pravin Paratey, CTO at Affectv a company that uses NLP and data-driven means to provided targeted and relevant ads to internet users.

We had an incredible turnout and the audience was fantastic. They were full of insightful questions and the discussion continued at the pub afterwards.  I suspect this will be the first of many of these events here at King’s.

Verifying Cloud Services

October 14, 2013 Leave a comment

What is needed to verify cloud services? How do we know whether the service we are using works as expected?

These questions are discussed in the recent article “Verifying cloud services: present and future”, by Sara Bouchenak, Gregory Chockler, Hana Chockler, Gabriela Gheorghe, Nuno Santos, and Alexander Shraer, published in the highly visible Operating Systems Review (OSR) journal (Hana Chockler is a Lecturer in the Department of Informatics of King’s College).

The team of experts addresses the challenges in verification of cloud services – from functional correctness to service availability and reliability, to performance and security guarantees. As cloud-based services become more and more popular, there is a real and growing need in tools for verifying these services. As the authors of the article argue, currently there is no adequate technology for verification of the cloud. They discuss recent research results that can help in bridging the gaps between what is needed and what currently exists in this area and suggest novel solutions.

The article, as it turns out, strikes a chord with the cloud users and service providers.  It recently featured in ZDNet (http://www.zdnet.com/do-you-believe-in-cloud-7000020607/ ), Forbes (http://www.forbes.com/sites/joemckendrick/2013/09/17/cloud-customers-are-not-getting-what-they-pay-for-study-says/), CloudTech ( http://www.cloudcomputing-news.net/news/2013/sep/18/new-paper-questions-whether-cloud-consumers-get-what-they-pay/ ), StorageMojo (http://storagemojo.com/2013/09/05/verifying-cloud-services/ ), and, most recently, BCloudReady.com (http://www.bcloudready.com/cloud-buyers-beware-make-sure-youre-getting-what-you-pay-for/ ), starting interesting and lively discussions about availability and reliability of a cloud.

Do you believe in cloud?” – asks ZDNet, listing specific areas, where, according to the OSR article, more tools and information are needed:

  • trusted software and service identity;
  • functional correctness;
  • performance and dependability;
  • security.

The Forbes blog entry starts with an eye-catching title “Cloud Customers Are Not Getting What They Pay For”, summarizing the takeaway from the article as “Cloud customers expect a certain level of service when they sign on to agreements. But a lot of the time, cloud services aren’t delivered as expected, and there isn’t even a way to verify that cloud services are performing as they should.

This paper is sobering because it shows how primitive current tools for verifying cloud services are – if they exist at all” – says StorageMojo, adding: “It isn’t even clear that cloud providers themselves have the tools to know the answers to questions that corporate users should and will have.”

BCloudReady .com recommends reading the full article: “… wading through the 14 page no nonsense report is worth the read. Compiled with contributions from six of the leading cloud experts worldwide it covers everything from verifying a strong service identity to protecting yourself from a “Byzantine Provider”. Well written and clear in both its methodology and conclusions it can serve as an excellent basis for evaluating your current cloud service provider(s) or as a guide to developing your strategy for utilizing cloud services.

For more information, see “Verifying cloud services: present and future “, by Sara Bouchenak, Gregory Chockler, Hana Chockler, Gabriela Gheorghe, Nuno Santos, and Alexander Shraer. Operating Systems Review, 47(2):6-19 (2013).

Big social data repository opened to everybody – Partnership between Department of Informatics at King’s College London and Affectv

Nowadays, we read more and more about big data and how they can be used by companies to their “business advantage”. For the last few years, companies gathered large datasets about their customers, employees, partners, etc. The challenge now is to analyze those data in a way that enables us to discover underlying patterns and trends, and from the business perspective to turn this knowledge into measurable profit.

It is not easy to mine meaningful information from an endless number of bits gathered in our computer systems. Things get even more complicated when we consider data that describes people and their activities as in such a case we have not only information about individuals but also their interactions with others. This means that we have to dig into a large graph that represents this web of connections.

Social targeting company Affectv is one of the businesses that understands the urgent need to look closer at the assets they have. Yes, data is an asset, a very valuable one. Thus, they decided to open up its aggregated social data for a limited period. The Open Graph Initiative kicks off with the Department of Informatics at King’s College London and will roll out to Edinburgh University and Indian Institute of Technology with the aim of sparking innovation within the social media interaction industry.

This partnership aims to help to uncover insights and trends in how social interaction across media platforms affects certain outcomes, for example:
• Can you identify interests based on someone’s social graph?
• Can you infer brand favourability through analysis of a person’s social graph?
• Can you predict purchase intent from social endorsement?
• What social signals reveal the strength of a social connection?

The Open Graph Initiative is organized in a form of challenge that is open to everybody who wants to be involved. It will consist of two open challenges per quarter. Each quarter, Affectv will release the results on its website. Credit will be cited for the relevant parties that uncover the most interesting insights on human behaviour on the social web and will receive press coverage.

So if you are interested in mining big social data – start now! Go to http://affectv.co.uk/opengraphinitiative or contact me (katarzyna.musia@kcl.ac.uk)

Related article: http://www.exchangewire.com/blog/2012/10/11/social-media-meets-big-data-affectv-announces-open-graph-rd-partnership-with-leading-universities/

Profile of Microsoft

August 20, 2012 Leave a comment

The August issue of Vanity Fair has a long and interesting profile of Microsoft and its management policies, Microsoft’s Lost Decade, by Kurt Eichenwald, available here.

Categories: Computer Industry
%d bloggers like this: