There’s no doubting that semantic apps are going to play a very important role in the future of the web so I’m always keen to speak to people who are positioning themselves and Australia to play a key role in that future.

One of those people is David Petar Novakovic and this is part 1 of a 2 part post I’m going to do on his work and his thoughts on how Australia is positioned to compete in the future of the web.

Queensland-based Novakovic has been working on a technology called Mango that is a semantic space technology. Mango uses “high dimensional representations of concepts, people, documents, or anything, to find relationships and connections in the vast amounts of data on the web”

The technology has an interesting history having its origins in a now defunct Co-operative Research Centre (CRC) based at QUT called the Distributed Systems Technology Centre (DSTC). It was later purchased by DistIP, a company owned by Queensland based ICT incubator, inQbator, which was created to commercialise research from DSTC. DistIP subsequently spun-off the technology into a new company called Comvine, which is where Peter is working at the moment.

InQbator says that: -

Comvine is a powerful contextual search engine and semiotics platform which searches on the meaning of phrases rather than on pure text. It captures the concepts that subscribers write about and connects parties, content and groups interested in the same topics. It will form part of the new breed of web applications in the Smart Web (Semantic Web/Web 3.0).

Interestingly, inQbator also lists another semantic technology under the DistIP heading called ShEBA which it calls “a powerful and disruptive patented decision-support technology” so keep an eye out for more info on that.

But back to Comvine, Mango and my interview with Novakovic.

—————-

Do you work by yourself or in a team? If a team who else is in it?

Recently Nathan Oorloff joined me on the team, we are working together on the code and scaling of the platform to deal with the load coming in from our partners. We are both really interested in large scale systems, so we are having a great time. I also have a great team of investors and advisors in inQbator, who fund my work and have been very supportive since day 1.

What got you into CompLing, semantics, natural language processing etc.?

Actually, my work did. I met Peter Bruza who was the information ecology project leader at DSTC through my work with trying to understand the technology. At the time I was looking around for something to do for my master’s degree and I think Peter saw that I was getting really interested in this stuff. He asked me to do my Masters by Research with him, so I did. As it turns out, Peter is a world renowned information retrieval researcher so I was really lucky to have him as my supervisor. It has been a very hard couple of years of full time work and full time research, but also has been very fulfilling. My work and research have both fed off each other, without either of them I would not be where I am now.

Can you explain, for the reader who has never heard of these fields, what they’re all about?

Generally most of these technologies target specific areas of what used to be known as Artificial Intelligence. AI became a bit of a cliche term, and a lot of the sub fields of AI rarely associate themselves with it any more. Generally machine learning, computational linguistics, natural language processing and information retrieval are fields that busy themselves with helping computers understand information in a way that enables them to help us humans do things better. Previously lots of systems used lots of rules to teach computers how to reason about things, but writing lots of rules doesn’t scale and is brittle (cat is an animal etc). These new breed of technologies are based on statistical methods that learn from real data, or high quality training data, no need to feed them manually written rules.

—————-

In Part 2 of this post we’ll look beyond Comvine and consider semantic apps, and Australia’s role in the future of the web, more closely…