Volkswagen’s dedicated R&D Hub uses data to enhance and improve automotive AI solutions.
Today, you can’t go far in the business or technology world without hearing the term AI.
Although there is still a debate on defining what Artificial Intelligence actually is, the concept of using machine learning to help make our lives more efficient and productive is a goal that many organizations continue to strive for.
In the automotive space, AI is being used to improve everything from driver safety to consumer convenience, to automotive sales and service. For OEMs, this represents a significant opportunity. At Volkswagen, applying AI to improve products and services is seen as a core strategy moving forward.
At VW’s U.S. operations in Auburn Hills, Mich., the automaker has created AI Detroit, a dedicated AI research and development unit in the Motor City. “Volkswagen has a tremendous treasure of data. Competencies in Artificial Intelligence will help us leverage this treasure, and in supporting people and business,” said Abdallah Shanti, Chief Information Officer for the Volkswagen Passenger Cars brand.
Daniel Weimer, head of AI Detroit, leads the team of machine learning scientists and software engineers dedicated to applying the latest AI and machine learning breakthroughs to the automotive business. It’s an unusual job in the auto industry, but one that Weimer and others at Volkswagen consider essential for developing, producing and selling the vehicles of the future.
Weimer says it’s really the extension of high-power data crunching that businesses have used for years. Today, that data and the specialized hardware needed to analyze it powers algorithms that can help improve results and get to the right answers faster. “We want to bring AI technology to business and create real impact,” said Weimer.
Weimer and his team monitor the latest developments in AI, from new techniques in machine learning to new software. They then work closely across Volkswagen Group of America’s different brands—including Volkswagen to Audi and Bentley—to find ways to implement those insights.
A key differentiator for AI Detroit lies in the ability to develop and operate AI solutions, feed them constantly with data from all across Volkswagen, and re-train the algorithms with new data. The next step will involve novel concepts that overcome the current limitations of how an AI is trained, an engineering architecture to build up even more powerful analyses.
“AI is a toolbox of technologies to create smart solutions. We’re trying to teach a computer to do things that are straightforward for humans, like understanding spoken language that can then be applied through software to real-world challenges,” said Weimer.
Weimer and his team are key to point that leveraging all this data does not mean using AI to replace employees with software. “It is key to develop technologies that serve and support Volkswagen employees,” Weimer said. “You let the AI do things AIs are good at, like finding patterns in huge data sets. But ultimately, our algorithms must always serve human decision-makers. After all, humans are better in strategic decisions, more innovative and more creative.”
One example of an AI Detroit application is a scheduling tool for workers at the Volkswagen manufacturing plant in Chattanooga, Tenn. With some 2,000 employees, and multiple shifts, internal rules and individual skill sets, the possible permutations of setting a schedule quickly become immense.
“We now offer algorithms to basically find a much more optimal schedule for management and workforce alike,” Weimer said. “We hope this can improve productivity, reduce physical stress for the workforce and even help them avoid sick days they may have needed in the past.”
Product quality improvements
Another project is the automated analysis and understanding of the text in hundreds of thousands of documents to further improve product quality. The team in Detroit has developed an AI-based Natural Language Processing (NLP) technology that analyzes reports and claims to check similarities and patterns. This systematic analysis can help Volkswagen engineers detect possible quality issues quickly and feed them back into the early stages of product development.
AI at Volkswagen has also been optimizing how to structure incentives programs and how to position different carlines in the market. By drawing on terabytes of sales data going back nearly 20 years, this system can reduce the guesswork of a model launch. Here again, AI algorithms support the human experts at the corporate level and dealers in decision-making.
As vehicles grow more digitally connected and demand more data for future technologies like automated driving, the need for AI solutions will increase. By locating its AI hub in Detroit, Volkswagen wants to attract researchers and developers who want to shape the future of the automotive industry and transportation—potentially worldwide.
“We’re in Detroit for a reason. You can knock on every door here and there’ll be someone with high automotive IQ,” Weimer said. “There’s all this awesome infrastructure and talent here, along with a great attitude of wanting to redefine the industry. There’s no other place where you find that.”