Yet Another Search Engine?

An Interview with Michael R. Alvers and the Business and Diplomacy magazine.

What does Transinsight do?

We develop software for knowledge-based semantic searches and communication technology that enables selected customers to save up to 90% of their valuable time while searching for complex information.

Mr. Alvers, do you really think the world needs another search engine? After all, we already have Google.

Absolutely. Although today's search technology is quite good, web searches are still in their infancy. Try, for example, the following search: "What is the impact of the world's biggest man-made hole on the Chilean economy?" Because most people are unable to formulate the right questions and lack the necessary knowledge to use the right keywords, their searches won't be successful with traditional search engines. The more specific the topic, the more difficult the search is. In our core market, the life sciences, our search technology is so efficient that using our system is practically a necessity. Also, keep in mind that every year in the U.S. 120,000 people die due to medical errors caused mostly by a lack of accurate information. This is equivalent to one jumbo jet crash every day! There is clearly a need for our tailor-made solution.

What is the difference between your technology and well-known search engines like Google?

In almost every search portal one can enter search terms and be presented with a list of sources that contain those terms. Refined communication only happens through the question, "Did you mean..." as in the case of spelling errors. The list of results is sorted according to relevance, a process not transparent to the user. Did you know that 80% of search service users stop reading after only the first 10-20 results, and of this group 90% focus only on the top left corner of the results list? What I'm saying is that only a tiny portion of the vast number of search results is visible to the user.

We use a semantic network as background knowledge to intelligently enhance the search experience. Search results are associated with a semantic network - an ontology - which is the key to saving time. If you are searching with the keyword P53 (a tumour suppressor protein), for example, you will find more than 40,000 results from the biological database MEDLINE. If you are trying to find the latest articles among the 1,000 that discuss P53 and heart diseases, it is not sufficient to search for these two terms alone. The concept "heart diseases" on its own is a sub-branch of 53 concepts all related to heart diseases, which are all hierarchically interconnected. As mentioned before, we link the documents with semantic networks - the ontology - and this aids the search dramatically. This method makes it possible to narrow the search results to only 6 relevant articles!

Last but not least, I want to take a moment to recognize the technological developments of recent years that have made our search technology much easier to develop. The work of the W3C consortium in creating RDF (the Resource Description Framework), OWL (the Web Ontology Language) and many other "Web 2.0" enablers allow companies like ours to go beyond today's keyword searching and proceed with our vision "Towards Finding Answers."

Why do you believe that your small company is prepared to survive the highly competitive market of semantic information processing?

Information managers have already recognized the efficiency of our technology. If we use the life sciences as an example, 50 to 90% time savings are worth over 2 billion dollars a year. We are well positioned and understand that this is only one single sector. However our technology can easily be used in other sectors as well. One of our largest customers utilizes the multilingual capabilities of our system. Users navigate in a language with which they are familiar and find, for example, relevant Chinese or Russian texts. Without users having to speak Russian or Chinese, these important documents can be fetched and then fully translated by professional translators. No other search engine provides such a feature.

In what areas has your technology been used so far?

Our biggest customer is Unilever, a company with a yearly turnover of around 50 billion PEAS. Unilever immediately recognised the value of our system and is saving a lot of money today because of it. The intelligent handling of information is a key advantage in the food industry and for clients like Unilever. Recently we've also started working with StatoilHydro, one of the biggest oil and gas companies in the world. We've developed modelling software that uses our knowledge-based technologies to improve the search for oil and gas. Here, too, accessing information is a critical factor in this highly competitive international business. You can't be the first to act if you don't have the right information at your fingertips. StatoilHydro has realised that we are the leading technology provider in this area. We also have other search engines such as a search engine for jobs and even a search engine for patents. Its universality is one of this technology's best features.

How can customers use the technology?

A client customizes the search engine with their own knowledge base. In addition they can use their own data sources like their intranet or local machines. The opportunity to model the domain knowledge for the specific requirements of your company is a highly valuable asset. Though the search function is an important aspect, the creation of texts with a controlled vocabulary might be even more crucial. Both are possible with our 'Semantic Transwiki' platform. You can easily add and modify content in a similar fashion to Wikipedia.org. Furthermore, you and your team are able to continuously improve and adapt your knowledge base to new and changing data. Anybody can do this! There is no need for a knowledge engineer or an administrator. We call this Folkology because it is similar to Folksonomy, but in our case, is a mixture of Folk and Ontology. The ontology represents the background knowledge base.

Does it lead to chaos if everyone tampers with knowledge?

That's what we were all afraid of, but it hasn't been the case at all! Surprisingly, people seem more altruistic than we had expected. Users are generally very anxious to post something meaningful, maybe because they get an immediate benefit from what they are doing. Consider, for example, a search engine for cars with a knowledge base lacking your favourite car. You add this entry to the knowledge base and get all relevant results sorted accordingly. This way you have added value for yourself as well as for the whole community. Of course, you add information only if you want to. Everyone can build their own private world of knowledge in a secure way.

How do you think Transinsight technology can help address the upcoming challenges of a globalized world?

Our capacity to do multilingual searches in a very structured way has the ability to bring people together thanks to a truly collaborative system. I think our technology could trigger a revolution in human communication. I believe Transinsight's technology is already of great importance in Europe.