The next frontier of Artificial Intelligence
Software makes automated decisions everywhere by using Artificial Intelligence (AI). Applications based on AI will redefine many industry products and services. Our mission today is to seize the opportunities offered by digital trends for industrial application areas and to turn them into business.
The current boundaries of broad application areas are now being tested by thought experiments in research and in the design of prototypes. Future developments in AI will be achieved at the same fast pace. There are two main reasons for this: technological progress and an expanding open innovation culture to further improve well-performing algorithms.
The accuracy of automated image recognition has leaped from about 70 percent in 2011 to more than 95 percent today, putting it on par with human beings. This tremendous progress is largely due to deep learning methods. Here, complex algorithms are used in multilayered neural nets that learn, on the basis of huge volumes of data in a training phase, create patterns on propositions and apply these new findings to new images. One strength lies in the simultaneity of mathematical operations, which are highly suited to deep learning tasks.
Currently, the capability of machine learning relies on large amounts of training data. But research into more efficient algorithms is in full swing. Predictive or real-time-learning algorithms will be the next big thing. Another accelerating factor of AI is the knowledge openly shared by large IT as well as Internet companies closely interacting with academia.