Sci-fi movies and TV shows have long covered the future of artificial intelligence including popular characters such as The Robot on Lost in Space, Pixar’s WALL-E and Data from Star Trek: The Next Generation. While Hollywood takes a fairly entertaining view of AI, the reality is there are many ways AI will enhance both our business and personal lives. In particular, great strides have been made in the usefulness of the artificial intelligence to enable human-like cognition with common devices by using machine learning. As recently as 2015, a deep learning neural network was able to successfully identify objects and animals in a picture more accurately than a human. Now the focus for machine learning (ML) centers around cloud computing due to the processing needs of ML. This is shifting rapidly to the edge for a number of reasons including the improvement in neural network model efficiency, the emergence of high-speed neural network accelerators, the need to process the huge amount of data being captured at the edge, alongside privacy concerns and the performance gain achieved by eliminating the latency of cloud access.
Because of this progress, NXP Semiconductors is able to create the first i.MX applications processor, the i.MX 8M Plus, with an integrated machine learning accelerator that can process neural networks roughly thirty times faster than Arm® processor cores. And in order to make this a complete solution, we are giving it eyes and ears by adding two camera ISP inputs and eight microphone inputs. Together, this creates a comprehensive ML solution for our customers that delivers compelling levels of convenience and productivity while maintaining user privacy. Imagine having a device that knows who you are, what you like, what you are trying to do and helps you accomplish more in less time.
With the i.MX 8M Plus, you can create a device that can pull up your personal presentation in a meeting room without having to be asked. It can understand what you are saying in order to know when to go to the next slide, or go backwards a previous slide. The system would know when to play a video clip or to look up information on the web to answer questions. It will be a personal assistant that helps you achieve more.
These capabilities will extend to your personal life in your smart home by turning on your fire place when you get home from a long day, playing your favorite music based on your mood at that moment or offering a selection of your favorite movies and TV shows without being asked. It could turn off your lights off as you leave rooms and turn down your heat when you leave on vacation saving money. In short, learning your routines to anticipate your needs.
Driven by your previous shopping habits and learning your needs, even retail will become more convenient. You won’t need a list, AI can make your list and help guide you through your shopping experience or even automatically order items you have preapproved. Cameras watch what items you put in your shopping cart matching them against your shopping list. Your cart will lead you to the location of each item you need and suggest alternatives if items are sold out. Your cart will suggest new things to try and how you can reduce your shopping bill. When you leave the store, you will not have to go through the hassle of a check-out line as your credit card will automatically be charged for your purchases. It will be a completely new kind of shopping experience and it is being developed right now.
Another example, “Where is my meeting? My party? My friend’s apartment?”. Imagine riding in your autonomous car to the smart city where traffic lights recognize your car and optimizes your travel time to your destination. When you get there, elevators will know what floor you need to go to without being told. After you are dropped off, your autonomous car will go park at a free lot outside the city or if enabled, will earn money for you by doing ridesharing while you are at work.
Machine learning is expected to also bring a new level of intuitive human interface to the smart factory by enabling image, voice and gesture control. The new i.MX 8M Plus has critical industrial features such as dual CAN-FD interfaces, time sensitive networking support with Ethernet and ECC.
These scenarios are achievable by leveraging the power of machine learning using specific algorithms such as face recognition, object recognition, speech recognition, prediction of intent and emotion detection. These skills and more will be available to run on the new i.MX 8M Plus and is why we created it. We provide a way for our customers to deliver these types of experiences which really excites me. These ML algorithms will make life much easier, more enjoyable and that is something we can all look forward to. Me and Hollywood included.
Learn More
Tony Masterson is CTO of AI and Media for Edge Processing at NXP. He founded the startup Zenverge, maker of video transcoder chips, that was successfully acquired by Freescale. Prior to Zenverge, he was VP of Engineering at multiple companies that were acquired including GlobespanVirata and iCompression. Tony Masterson received his MSEE and BSEE from MIT and holds 30 U.S. patents in the areas of digital video, computer graphics and architecture.
