Now that we are about to finish this 2016, it’s the appropriate moment to look back and analyzed what has happened. Watching the stats, we can affirm that one of the trendiest words of the year has been: artificial Intelligence, and the best is yet to come.
According to Forrester’s newest report the investment in Artificial Intelligence is expected to grow 300% in 2017 compared with 2016 across all businesses. This impressive rate make us thing that we will keep hearing even more about AI the upcoming year.
However, among all the applications of Artificial Intelligence, there are three that have captured our attention and will be growing impressively next year:
-Self driving cars: have been around for a while since Google and Tesla decided to launch their own, however after a couple of accidents there have been quite some doubts about their usability and safety. The growth and the embrace of self driving cars will be the perfect example on how much humans can trust artificial intelligence, and that is why we decided to include them on the list.
A key factor to start building trust is showing reliable proofs of how safe these cars really are, and the best way to prove it is testing them. Earlier this year, Google claimed that they have been testing their cars for more than 1.5 million miles.
But in our opinion, artificial intelligence is something that sounds like science-fiction movies to the biggest percentage of the population, so another key for the adoption of these cars is to explain the technology behind them, how they really work.
The underlying idea, is that the car system is equipped with a machine learning algorithm that collect data on everything that may affect the driving, such as human actions, weather conditions, traffic signs. Based on all this information the algorithms can predict how all this factors will combine and then act taking the best decision according to the given conditions.
But, what about the unexpected signals? Like the information written in the luminous traffic signs warning about an accident? This is a field inside self driving cars that needs development because traffic conditions change every day.
-Conversational robots: the idea of interacting with a company is not something new, it has been around for a while but every year it keeps evolving and in 2017 we are going to see many companies adopting some trends that started in 2016.
Conversational bots are one of these trends. They will allow companies to enhance their user experience while looking for products or services on their sites, and act on some of their pain points no matter the field: banking, retail, or hospitality.
However, as we have mentioned in previous posts, to truly enhance user experience it is necessary to increase the level of accuracy.
Machine learning has been able to provide very good results, however the process to train bots using this system is quite slow since to be able to learn properly it requires a large amount of time, but in highly competitive markets time is money.
In 2017 we believe we are going to experience a speed up of this training process without decreasing the accuracy level by using the appropriate training methods and materials.
We recommend you to watch this webinar where we explain the differences between the actual system and the upcoming one and how beneficial it can be.
-Cybersecurity: in November 2016, some of the biggest companies in the world such as 3M or Deutsche Telekom suffered cyber-attacks. But they are not the only ones, according to data on 2015 there were about 200 cyber-attacks per hour that left behind loses for about $400 billion.
With the proliferation of services on internet, quitting online resources is not an option, so in our opinion the solution is based on prevention. Our research made us realize that there is not much we can do to prevent cybercrimes, so this is a major are of improvement for 2017.
How is possible to prevent cybercrime? By monitoring and extracting information of the sources criminals use to communicate: dark forums and social media. This means, however, a lot of work and it’s almost impossible to do manually.
Technology can help to improve the monitoring and analysis of all this information by using a PDA-based non-deterministic GLR parser to analyze text and that allow us to extract the main concepts and a lemmatizer to collapse all the forms of a verb to its root.
By extracting automatically, the main keywords we can know about the tools criminals use to attack, how, and which resources do they need. But the key here is to have a tool based on linguistics, able to solve the problem of ambiguity and to distinguish the different meanings of the words.
If you want to see some examples and see how powerful our tool is to extract data from different types of text sources?