Artificial Intelligence (AI) leading to real-world possibilities
Posted: September 11, 2019 by John G. Smith
INDIANAPOLIS, Ind. – There’s data, data everywhere, but limits to what people can do with it. Advances in Artificial Intelligence – often referred to simply as AI – is introducing the computing power that will harness the data to guide everything from the decisions made by autonomous vehicles, to the efforts to better manage traffic flows.
Advances in AI are made possible by today’s ability to capture immense volumes of data, exponential increases in computing power through cloud computing, and gains in machine learning, explains Alexandre Barsi Lopez of the Indiana University Kelley School of Business.
“We can do things that were not possible before even a couple of years ago.”
Sensors are emerging in areas that would never have been considered before, he added, during a presentation at FTR’s annual conference. And not just on the newest systems. During a recent visit to a production line, he saw sensors monitoring equipment that was still 25 years old, feeding instructions to personnel by cell phones. “You don’t need to have the most modern technology to add sensors everywhere.”
The advances are expected to play a key role in security, too. “It’s impossible to think about every single attack that people can come up with, but what you can do is say what is the current state of your data, what is your current state of infrastructure, and you can start looking for variations,” Lopez said.
New approaches for handling data, and changing the amount of data needed to determine a pattern, could make AI cheaper and more efficient, he said. The same can be said for approaches that involve offering AI as a service. Business systems like Salesforce and SAP already offer AI add-ons to their offerings.
The computing power won’t necessarily replace people. It could be used to limit repetitive work, such as taking the first look at hundreds of job applications. Other work could run in parallel, with a machine completing some work alongside a human. While someone takes a phone call, for example, a system could use language processing to listen in to the conversation and predict the tools and parts that are needed for a service activity, he said.
Even cities are becoming smarter, and Lopez believes 5G networks will accelerate the possibilities of AI. Most public garages in Beijing are now connected, displaying the available parking spots whenever a car enters or leaves, he said as an example. Imaging the applications of knowing exactly where parking will be available.
High-speed networks – especially in cities – will improve access to vehicle data and the ability to make remote adjustments, he added.
It could be a key to winning last mile deliveries, too. Kroger stores are testing autonomous delivery bots in the U.S., and Amazon is testing drones. “There a lot of creativity right now,” Lopez said of the tests. Combine it all with machine learning, and routing could be improved to manipulate traffic flows.
None of this will involve plugging and playing, though.
“AI is not something we can say, ‘Oh, artificial intelligence. Here it is,” said FTR CEO Eric Starks. Data can be shared in different directions, or even in both directions, but different tasks will require the data more often.
Autonomous trucks offer another layer of challenges.
“What happens if the connectivity gets lost? What happens if it’s not sending the right data? How much duplicate information do we have to have? All these things need to happen,” Starks said in a briefing of his own. Even batteries will need to be considered. “It eats up power. When you process stuff, that needs power.”
With the growing interest in AI, decisions need to be made about storage space, likely relying on cloud-based computing. “As we continue to move down this pathway, is that available and enough?” he asked. “The amount of processing power that it takes is significant to start doing things like AI, to be able to process all this information.”
“We have to be able to share this data so that it becomes useful, so a new tool can be devised that allows us to do a better decision.”