Semantic search engines: reality or fiction

For some years now, we have been talking about semantic search, Web 3.0 and a list of concepts related to the semantic world and the idea of people finding what they are searching for.

Any new technology able to defeat Google is of great interest to all of us, specially when it takes advantage of famous semantic technologies. However, it seems that by now the results are not that different, so they don't impress very much.

For instance, when we search in Google: ‘when was Leonardo da Vinci born’, the answer is clean: 15th April 1452. And if we keep searching, rephrasing and abridging the search, ‘Leonardo da Vinci’s birth’, similar results are offered using natural language. We are doing something wrong…

Let’s see what may be happening:

Starting from the handicap that semantic search engines tend to be presented in a search box similar to the one in Google and allow searches with the format we choose. For this reason, we write what we are used to ask: primitive queries.

Moreover, as I recently read in a quite prestigious blog, the first confusion in this area is that semantic search engines are increasingly used as an answer to all the possible problems, from modern searches, currently controlled by Google, to problems impossible to solve from the computational point of view.

Por lo tanto, creo que en este sentido, se debe tener muy claro cuál es el objetivo que queremos cumplir con este tipo de tecnologías y, sobre todo, se debe acotar muy bien el conocimiento sobre el que el motor semántico va a trabajar y trabajar de acuerdo con los contenidos y la forma que tienen los usuarios de buscar esta información.

For this reason, I believe we must have a clear purpose to deal with this type of technologies and set specific boundaries to the knowledge and contents the search engine is going to work with, and to the ways the users use to find this information.

A good approach to deal with this type of web, for instance, would be to combine both technologies in order to cover searches in natural language and also traditional searches. One successful case in this sense is, for instance, the Cajasol website (https://www.cajasol.es), where you can make queries like:

  • ‘tarjetas cajasol’ (cajasol cards)
  • ‘hipotecas’ (mortgages)

or else queries in natural language:

  • ‘como puedo enviar dinero a colombia’ (i would like to send money to colombia)
  • ‘acabo de perder la visa y quiero bloquearla’ (i just lost my visa and i’d like to block it)

Conclusion:

I believe semantic search is a technology that is creating a lot of expectations and in the future it will be able to solve problems that Google can’t currently solve.

Inbenta Team
by Inbenta Team