Artificial intelligence applications are proliferating at a rapid rate. The global market for AI and cognitive systems is already worth $12.5 billion, reflecting a growth of 59.3 percent since last year, International Data Corporation estimates. This will continue to increase at a compound annual growth rate of 54.4 percent over the next few years to reach $46 billion by 2020. Quality management, medical diagnostics, automated customer service, security and fraud investigation will be some of the application areas that will see the highest growth rates.
Most new innovation in the field will take place in software applications, because AI hardware development has already been consolidated by industry leaders. Here’s a look at three artificial intelligence hardware and software innovations that illustrate how rapidly AI has matured and how sophisticated today’s technology already is.
Mobile On-Device AI
The most significant recent development in AI hardware is the emergence of mobile device components that can support on-device artificial intelligence applications. Until recently, the amount of computing power required for practical AI applications required either a PC with sufficient power or access to remote cloud resources, limiting mobile applications to the rate at which data could be downloaded. Now, however, as mobile manufacturers prepare to handle 5G download speeds, smartphone processors have gotten fast enough to handle on-device AI.
For example, Qualcomm has recently introduced its Artificial Intelligence platform, designed to be capable of running on-device AI applications. The platform’s design empowers smartphones to use AI machine learning for applications such as optimizing battery conservation, smart camera focusing, streaming virtual reality, filtering out background noise and biometric user authentication.
Qualcomm’s artificial intelligence platform also supports facial recognition, another already-popular mobile application of AI. One of the companies at the forefront of mobile facial recognition is Snapchat, which uses AI to create selfie filters. To fit a filter to your face, Snapchat’s app employs a database of thousands of faces that have been manually marked with points that represent patterns typical of human faces. By applying machine learning and computer vision, Snapchat can match facial patterns from its database to the pattern of your face, creating a filter that shifts in sync with your facial movements.
In order to get facial recognition to work on mobile devices without causing performance lags, Snapchat had to figure out how to design its app to put AI on your phone. To achieve this, Snapchat simplified and modified the way its neural network technology works. The result is a facial recognition algorithm that only takes up 5.2 MB, about the same amount of space as an MP3 song.
When most people use ridesharing apps such as Uber and Lyft, they probably don’t realize they’re using an AI application. Artificial intelligence enables ridesharing apps to minimize your wait time when you hail a car, match you with other passengers to avoid detours and set your price.
Uber originally turned to AI after it had outgrown its original vision as a way to request premium black cars, expanding to encompass a range of ride services and to handle millions of rides a day. To manage this expanded growth, Uber redesigned its app by applying AI machine learning. Machine learning’s ability to process data rapidly enables Uber’s redesigned app to perform tasks such as calculating the shortest route to get to a destination, predicting rider needs, recommending optimal pickup points and avoiding layovers for commuters using Uber in conjunction with public transportation.
On-device artificial intelligence, facial recognition and ridesharing apps illustrate some of the cutting-edge hardware and software innovations in AI that are already in use today. As AI applications continue to proliferate, they will increasingly permeate everyday consumer and business experience, transforming the way we work, shop and live.