What exactly is artificial intelligence about?
AI revolves around the development of complex algorithms enabling computers to be trained to perform specific human tasks. This could be the automatic analysis of images or recognizing, interpreting and translating texts. But the technology can also be used to intelligently automate various operational company decisions, such ad the detection of tax fraud. The algorithms may not be equipped with the human capacity to innovate or to reason but very often they are faster and more accurate in executing very specific, very well delineated tasks.
From cars over medical imaging to industrial applications
The past few years, companies such as Google and Tesla have been pioneers in introducing AI with a large audience, showcasing applications such as Google Translate or Tesla’s Autopilot. At SAS, we believe it is one of our missions to expand AI’s potential towards the entire corporate world as well as the non-profit sector. At this moment, we are working towards that mission by putting our SAS VIYA platform to use in a number of selected areas. One of our priorities is in the analysis of medical imaging material. RX images are an important source of information in the process of diagnosticating. But we are currently facing a huge challenge in processing the increasing volume of medical imaging efficiently and accurately. Artificial intelligence can lend a hand here, in recognizing specific patterns in labeled biomedical images. Algorithms may prove useful here, e.g. in formulating predictions or in classifying images, thus supporting the - often overburdened - radiologists.
The explosion of data in the industrial world due to new technologies such as the ‘Internet of Thing’ paves the way for many industrial AI applications. One of our customers in this area is a world-leading computer chip manufacturer, who actively deploy 3D technology featuring a fault tolerance of one in a billion. This may seem safe enough, but at the current speed of 50 million drops per second, this means you may end up with one mistake every 20 seconds. AI algorithms enable an automatic analysis of photographic images of the production lines. This allows for a timely detection and reduction of errors.
Industries featuring a high degree of so-called “process variability”, such as the chemical sector, can benefit from analyses of sensor data measuring external circumstances such as temperature and air pressure. Machine learning algorithms will then enable a better insight into the relationship between these data and the process parameter settings and productivity. The control parameters can thus be optimized in real time, in order to realize a significant increase in efficiency. Our projects have led to our customers increasing the output on their production lines by 20 %, at the same time reducing their energy spend significantly.