Looking back on AI events in September 2017
September 2017 was full of events that explored the future of artificial intelligence and made the case for getting involved in its ongoing development.
IntelliSys, which took place on Sept 7th and 8th in London, offered the opportunity to delve into the latest developments around AI and machine learning. Valentina Salapura, who’s working with the IBM Watson team, gave the opening keynote, taking her audience on a stroll through the history and various applications of new AI-related computing technologies. She drove home the point how our notion of a computer needs to change going forward -- no longer perceiving it as a machine to feed data into and then expecting it to simply execute commands, but rather a device that can tie together all available data to arrive at the most intelligent decisions. Machines of the future will offer solutions since they can deal with massive complexity far better than the human brain -- or a regular PC, for that matter.
Frank Wang from the University of Kent in Southern England introduced the fundamental research of the discipline.
The third keynote came from Jan Hofmann with Deutsche Telekom. He gave a hands-on tour of his company’s experience with leveraging AI for customer care. For starters, he sketched the challenges such as a high volume of requests, often repetitive topics, and the overarching goal to minimize costs. To tackle those challenges, Telekom built a chatbot called TINKA.
It’s initially being tested in Austria and will later be introduced to the larger German market. TINKA is able to learn, can communicate and have a dialog much like a real person. Its limitations notwithstanding, TINKA is already seeing higher acceptance among customers that an earlier, purely rules-based assistant. There’s room for improvement, Hofmann said, because the technology on the market has to catch up. Even though everybody expects those systems to be able to learn effortlessly, they are aren’t yet. The necessary tools should be easily available, but they’re still in their infancy. While there is plenty of data to feed into the system, the question remains what data is the right data. And even though language processing capabilities for dialog systems have been around for awhile, they still need some polishing to truly shine.
In the future, AI will be able to learn more efficiently since processes will adapt to the developments in the telco market without the need to be hand-fed. There are also efforts underway to localize across multiple languages, which is important for geographies such as the Eastern European market. The guiding principle for all this development work ought to be to keep the technologies open and interoperable.
Paolo Rosso from the Technical University of Valencia and Nuria Oliver with Vodafone took the stage on day two. Rosso demonstrated how language can be analyzed to recognize misleading information (e.g. fake reviews on a website), irony or sarcasm. Nuria Oliver spoke about her work on collecting and analyzing smartphone data for the purpose of crime prevention.
Go Ignite, Berlin
There were a total of five awards given away during Start-Up Night in Berlin on September 8th, all presented by Go Ignite, the international alliance of telco heavyweights Deutsche Telekom, Orange, Singtel and Telefonica.
The following week, on September 11th and 12th, the winners of an open call for startups in Madrid could participate in a workshop. Five startups were recognized for their innovative solutions in three technology areas: AI in customer care, connected home and cybersecurity for the IoT. The winners were Sparkcognition and NanoLock Security for IoT security, Cujo and Vayyar Imaging for the connected home, and SafeToNet in the area of AI-powered customer care. All five companies had the opportunity to get a better understanding of the 1.2 billion people-strong customer base of the four telco partners behind Go Ignite.
RPA&AI Exchange, London
Senior-level executives only met in London on September 27th and 28th for the exclusive RPA&AI Exchange. Dan Fiehn, Chief Information Officer of insurance company Markerstudy, gave the keynote. All talks during the event revolved around the machine-driven future and, more precisely, robotic process automation (RPA). The big question here: How best to develop, deploy and implement intelligent, AI-powered automation? The second day shifted the focus slightly since two telco executives took the stage.
Gourav Datta, Senior Transformation Leader with BT (British Telecom) spoke about the AI experience inside the telco, while Jan Morgenthal with Deutsche Telekom provided insights into how his company is using AI for customer care. The final word belonged to Leslie Willcocks, professor at the London School of Economics and Political Science, who weighed the risks and challenges of developing RPA going forward.