Semantics is Not the Magic Potion, but it’s a Key Ingredient
With the launch of our newest white paper “Semantic Technology: What it Means and Why it Matters” we interviewed one of the creators of our flagship Cogito semantic technology, Marco Giorgini. In addition to his role as head of R&D, Marco is a writer of short stories and creator of video games and mobile apps. He is based in Expert System’s Modena, Italy office. This is the first of three posts.
What can we expect from the semantic technology of the future?
Marco Giorgini: When I consider the possible evolution of the semantic approach, most of my ideas seem to be taken from sci-fi stories, but, well, our present has already gone further than science fiction predicted less than 50 years ago, hasn’t it? And the fact is that we all are working to make the difference between a real automatic comprehension, and a pragmatic simulation of it, smaller and smaller. So, even if I won’t say I imagine a world with intelligent computers—because that would seems a bit scary—I really imagine a world in which a computer system can read all the information it finds (with read I mean process, which is to say, classify, organize and store, like we do, but with no memory holes, and with better speed for connecting information, even if with no hunches or whatever it that makes us clever, or artistic). And if that comes true, or, better, when it will come true, likewise will follow tools with support for everyday tasks, to retrieve and provide us hints of info to help us to do things faster, and with a better view of them. It would mean that our awareness of things will be enormously improved (if we wish), and that should help us plan our tasks better.
These specific concepts are my vision: Sentiment analysis (of any kind), and data summarization. For example, I expect that I could ask “what’s the preferred red wine for a tagliata di manzo” and get a single answer taken from all the available sources –social, blogs, newspapers, etc. I would also expect to be able to expand that further to discover statistical distribution, and even further, if I want and if I have the time, to actually point to the best source that tells me what I want to hear. Ranking, relevance and everything else could be handled in the background, and I could be helped to select things I want according to a whole range of tastes, in a matter of seconds (and those are the key points: smart highlights from big data, in seconds).
Another example: I could ask if it’s wise to visit Chicago this November and get, like before, a smart, multilevel answer, like yes/no, and expand that with a why?, and actually get an explanation, with the ability to get even more details.
Here, the answer could deal not only with people’s reviews on travel sites, social, blogs, but even with local events (referencing transportation strikes, festivals, or so on), traffic and weather statistics, the current political/social/criminal situation (if the timeframe is close) or statistically in that specific period. It could even pull information from my Twitter posts or my email (if I so choose) to incorporate my personal preferences.
Semantics is not the magic potion—but it is the whole set of magic ingredients that make up the magic potion. It provides and it will provide better and better info from unstructured data. And once it will be good/fast/precise enough (it is already, for a lot of things and data volumes), and it will be possible to mix everything up in a richer and more complex way, and all will be commonly available, we won’t even understand how there it could have been a world before that—like the odd one before the internet.
Connect with Marco on Twitter @marcogiorgini