9 April 2019

Convergence (The Acceleration of Ai, ML, NLP, and DL)

Convergence used to be a word with a great amount of hype in 1999. The two technologies that were rapidly involving (the mobile phone and the internet) were expected to come together at some point in the future. And that is exactly what happened.

Most of the experts had the view that this would yield amazing results. Turns out, that they were not wrong after all. The internet had come a very long way. It started with dial-up modems, after which way was given to DSL. Perhaps the most interesting thing about the prediction of mobile phone and internet convergence was that neither of these technologies was fully understood at that point.

Each of these technologies at that point on time simply reflected a range of different possibilities. Today, we already know the large impact that the internet has caused. Moreover, delivery over the mobile phone (something which started off via the internet) is now a norm.

Today, we again stand at a very similar place again with respect to the convergence of artificial intelligence with various different technologies.

The idea of artificial intelligence has been ideated from more than 50 years. However, recently, artificial intelligence went through a number of breakthroughs. Consider Watson for instance, a system developed by IBM that makes use of machine learning, big data, and natural language processing.

Back in 2011, this system otherwise dubbed as Watson won a Jeopardy game against the two top champions. This was simply a commercial reflection of how powerful artificial intelligence can be. According to Gold, the prime aim was to prove that various technologies can be brought together in very innovative ways so as to perform tasks that are seemingly impossible for machines.

Similarly, now, Google and Siri have redefined the way human-data interaction occurs. Over the past years, systems such as Google and Siri demonstrated that it is not necessary to be hooked on to a laptop in order to interact with data.

Now, in this model, artificial intelligence systems are making a transition from speech recognition, to the natural language interaction, finally moving on to natural language generation. Eventually, these systems will become capable of writing, and receiving information. While these may sound small, in reality, these are big breakthroughs that will have a great impact in the future.

Where on one hand, machine learning requires the assistance of humans, deep learning systems are capable of learning on their own without the need of humans. Previously, systems were simply trained to do so. Such a convergence of artificial intelligence and cognitive computing means that machines will no longer need the aid of humans to learn, in the future.

The Turing Test was created by Alan Turing in order to determine if a machine can exhibit intelligence behavior that is no different from humans. Now it looks like, within the next ten years, machines will be able to pass this test. Hence, they will fully be considered artificial intelligence.

Back to top