Apple Inc. has been expanding its artificial intelligence expertise with hiring and acquisitions. The company is working hard to keep up with competitors across mobile computing, which is growing to include smart watches, digital assistants, and autonomous vehicles. Apple’s latest purchases include two startups working on speech and image recognition technologies, which will likely become part of its user interfaces.
“In the past, Apple has not been at the vanguard of machine learning and cutting-edge artificial intelligence work, but that is rapidly changing,” said Oren Etzioni, a University of Washington professor and CEO of the Allen Institute for Artificial Intelligence. “They are after the best and the brightest, just like everybody else.”
Apple acquires AI firm to improve Siri
Apple bought VocalIQ Ltd., whose machine-learning technology is intended to help machines have more natural conversations with humans and facilitate hands-free communications while driving.
“The use of speech to interact with machines has reached a tipping point,” said VocalIQ chairman Steve Young. “Without smart conversational interfaces which can adapt to suit the user, the Internet of Things cannot flourish. VocalIQ intends to be the prime supplier of these smart conversational interfaces.”
Apple is expected to use VocalIQ’s technology for its Siri virtual assistant and possibly its electric car, since the Cambridge, U.K.-based startup worked with General Motors last year to develop voice interfaces to allow drivers to use speech to turn on windshield wipers or adjust stereo settings.
“Traditional spoken-dialogue interfaces are handcrafted, fragile and frustrating,” Apple said. “It is unrealistic to expect 7 billion people to start talking to machines in a way mandated by a programmer. Dialogue systems need to learn how people speak, and not the other way round.”
VocalIQ was spun out of the University of Cambridge’s Dialogue Systems Group, and it raised £750,000 ($1.1 million) in funding last year led by Amadeus Capital Partners. Apple declined to say how much it spent on the company.
Apple invests in image recognition, privacy
Apple also bought startup Perceptio, which has been developing technology to let mobile devices such as the iPhone recognize images without needing to share much user data. Terms of the deal weren’t disclosed.
More on Apple, Google, and AI:
Since Apple plans to restrict the amount of personal information its systems share in the cloud, it won’t be able to aggregate data for predictive analysis. So, for users to be able to get personalized services, the company is likely developing other ways of processing speech and image recognition.
Apple patents action camera
In January, Apple Inc. has received 34 patents, including one for a waterproof digital camera that is similar to GoPro’s video cameras, which are used by sports fans and drone hobbyists. Apple’s patent describes cameras that could be controlled remotely via a smartphone or watch.
Some industry observers have suggested that Apple buy GoPro because of its content ecosystem, but others disagree, and GoPro CEO Nick Woodman does not appear to be interested, and his company is working on a quadcopter of its own.
Apple admits to automotive ambitions
Apple’s VocalIQ and Perceptio purchases demonstrate how aggressively the company is racing into the self-driving vehicle space, with plans for electric and driver-assist vehicles by 2019, according to The Wall Street Journal.
Existing automakers have dismissed Apple as an upstart that doesn’t understand their industry, but given Apple’s success in making popular mobile devices a lifestyle signifier, its entry into the car market is no less improbable than that of search engine company Google Inc., now known by its Alphabet umbrella.
Apple has also tripled the size of its Project Titan team and discussed autonomous vehicle testing with California regulators.
In other car-related moves, automakers are integrating Apple’s CarPlay, which transfers iPhone displays to car dashboards. Apple also recently spent $25 million on Mapsense, a data analytics and visualization startup in San Francisco.
Apple has claimed that its vehicles will feature driver-assist technology rather than be fully autonomous such as Google’s planned self-driving car. However, AI research is still necessary for maintaining competitive advantage in the increasingly crowded, if not yet mature, self-driving market.
AI has many suitors
Apple is hardly alone in pursuing AI. Its Siri faces competition from Amazon’s Echo, Facebook’s M, Google’s Now, Microsoft’s Cortana, and IPsoft’s Amelia, not to mention emerging social robots such as Jibo and Aldebaran’s Pepper.
Early this year, Facebook acquired Wit.ai, a Palo Alto, Calif.-based language-processing startup whose application programming interfaces (APIs) can process voice commands and texts.
San Francisco-based Salesforce.com acquired Tempo, whose calendar app uses machine learning to integrate with e-mail and social media to look up data about upcoming events. Tempo was based on research at nonprofit SRI International, which also helped develop Apple’s Siri.
Google spends on deep learning
Not to be outdone, Google has invested in the German Research Center for Artificial Intelligence (DFKI). The DFKI has about 450 researchers, a budget of 41 million euros ($46 million), and numerous commercial spinoffs. Other partners include Airbus, BMW, Intel, and Microsoft.
“The global market presence of Google opens up a unique opportunity,” said DFKI CEO Wolfgang Wahlster. “I am sure that we can revolutionize mobility, living and working for the future with the next generation of autonomous assistance systems in innovation alliances with Google and our other shareholders.”
Google had previously spent $400 million on London-based DeepMind and invested in Oxford University’s AI research, and it is considered to be a leader in image and object recognition. Google’s Android M mobile operating system will support the voice-activated Google Now.
On the hardware side, Google is in a seven-year partnership with the National Aeronautics and Space Administration, D-Wave Systems Inc., and the Universities Space Research Association in the Quantum Artificial Intelligence Lab. The lab’s D-Wave 2X operates at 15 millikelvin and is being tested for machine learning.
Also, Intel Corp. has invested $50 million in quantum computing research at the Delft University of Technology. Experts expect quantum computing to be capable of handling calculations too complex for today’s supercomputers, including speech recognition.