KAISERSLAUTERN, Germany — Believe it or not, artificial intelligence (AI) research began in the 1950s by the likes of MIT professor Marvin Minsky who taught computers to play (and win) at games like checkers, to speak their minds with voice synthesizers and to perform other tasks deemed amazing in that day. However, by the 1970s AI fell into disrepute due to lack of progress, only to be resurrected in the 1980s in the form of “expert systems” which went through several generations ending in the 5th Generation Japan project that got the whole world re-interested in AI. That's when the German Research Center for Artificial Intelligence (DFKI) was founded.
AI again fell into disrepute, due to limitations on the “expert system” paradigm, but the German's continued anyway giving DFKI funding starting in 1988 and eventually set up the organization as a self-sustaining non-profit incubator for nearby technical universities. Today it continues a wide variety of AI research backed by an all-professor management plus 740 employees in its Kaiserslautern research center, including 420 post-doctoral researchers and 320 doctoral candidates.
It is funded now by contract research projects it performs for customers like Bosch, Google, Hewlett Packard, IBM, Philips, Siemens and many others. The staff members often move on to become professors, founders of startups or employees of the technology companies spread across Germany, according to professor Andreas Dengel, a senior member of all-professor management team at DFKI.
Dengel's speciality is decision support systems that identify what is relevant to people, how to make educated guesses about the future interests of people and combing through heterogeneous information sources to try and meet those needs–sometimes before the person themselves realize they have that need.
“We started by eye scanning to see what text passages people read carefully and which they skip over,” Dengel told us. “We've also experimented with adding sounds to text passages as they are read such as producing a growl when reading a passage about wolfs.”
Dengel's group also provides optional details about what a person is reading, such as background information about objects being read about in balloons that can be read or dismissed at the users behest.
His group is also anticipating people's interest from historical data about recurring events. For instance, each year the Emmy Awards causes a great spike of interest on social media outlets. To describe these interest areas Dengel and associates are using the resource description framework (RDF) to create semantic dictionaries that tie together the meanings of words with associated material that is relevant to human understanding.
Today (DFKI) is also applying its expertise in AI to application areas like the smart factory, according to professor Detlef Zuehlke, the director of innovation factory systems at DFKI.
“Now is the right time for smart technologies to be used in our factories,” Zuehlke told us. “We call it the fourth generation of the industrial revolution, with the first being the invention of the steam engine, the second being the invention of the conveyor belt, the third being the use of electronics automation and now the fourth being the introduction of cyber-physical systems [CPSs].
The fourth industrial generation is already adding hundreds of thousands of microprocessor-controlled equipment modules to the factory floor, all connected, and coordinating their use with data processing techniques that can process the terabytes of data they produce from anywhere on the internet, Zuehlke explained.
To read the rest of this article, visit EBN sister site EE Times.