Robots must be smarter if they’re going to pack boxes in warehouses, scan inventory in stores, and even care for the elderly.
The rise of machine learning in recent years is making that possible. Steady innovation has led to robots that can independently “learn” to navigate tight corridors and grasp delicate objects without crushing them.
Some of the leading American and Japanese robotics companies and investors recently gathered in Menlo Park, Calif. to discuss artificial intelligence in robotics and its impact on business. Their conclusion? The robots are coming. But it may require some cooperation between the U.S. and an important overseas ally.
Japan has long been a powerhouse in robotics, fueled by its huge appetite for automating its automotive and manufacturing industries. But these kinds of old-school robots lack the cutting-edge machine-learning software that could help them continuously improve on the job.
Recognizing the problem, Japanese robotics companies, in some cases, are partnering with U.S. companies that specialize in A.I. software. The goal is to take the best from both countries.
“The Bay Area is good at software engineering and Japan is good at hardware engineering,” said Tomochika Uyama, the Consul General of Japan in San Francisco. The combination makes for “a perfect pairing,” he added.
For example, IHI Corporation, a Tokyo-based industrial equipment giant, recently partnered with San Francisco startup OSARO, which creates reinforcement learning software for robots that helps them, through trial and error, pick up objects they’re never encountered before. The spawn of the partnership is now being used in IHI’s warehouses to grab objects like water bottles and toothpaste and then place them in bins on conveyor belts.
“We are hardware manufacturer, so we are looking for a partner who can bring us the new power of A.I.,” said IHI vice president Kohei Taya.
Amazon, which uses thousands of robots in its warehouses, is somewhat of a role model in robotics. The company acquired robot maker Kiva Systems in 2012 for $775 million, and has since used its technology to jumpstart its use of robots internally. Last week, during an event in Las Vegas, Amazon showed off two new robots that can more efficiently move and sort packages than older models. The machines are intended to speed up the warehouse work while saving money.
Now, it appears more robotic companies are trying to create the next big Kiva Systems that could help more businesses than just Amazon package goods and track inventory. If the recent robotics event in Menlo Park is any indication, some of these next-generation robots could be the result of U.S.-Japanese teamwork.
EYE ON A.I. NEWS
Twitter’s A.I. push to clean content. Twitter said that it acquired machine-learning startup Fabula AI for an undisclosed amount. Twitter chief technology officer Parag Agrawal said in a blog post that the London startup would help Twitter “improve the health of the conversation, with expanding applications to stop spam and abuse and other strategic priorities in the future.”
Robotic pedigree. Prominent roboticist Rodney Brooks and A.I. expert Gary Marcus have founded a new company, Robust.AI, that specializes in creating “smarter” robots in areas including construction and deliveries, reported TechCrunch. Brooks, who co-founded Roomba vacuum maker iRobot and the recently shuttered Rethink Robotics, will be Robust.AI’s chief technology officer while Marcus, who sold his startup, Geometric Intelligence to Uber in 2016, will be CEO.
U.K. military bets on A.I. The United Kingdom’s Royal Navy and Marines said it has partnered with Anduril Industries, a defense technology startup specializing in machine learning, to “modernize surveillance systems and techniques by utilizing cutting edge technologies.” Anduril Industries was founded by Palmer Luckey, who helped create the company behind the Oculus Rift virtual reality headset, which Facebook bought in 2014 for $3 billion.
HSBC sets sail to Canada. HSBC has opened a research and development lab in Toronto that will focus on artificial intelligence, making yet another company to debut an A.I. research lab in Canada, reported The Globe and Mail. The paper noted that the new Canadian outpost is the banking giant’s second research lab outside of London.
DON'T HOLD YOUR BREATH
IBM researchers participated in a Reddit AMA (ask me anything) session, and fielded questions from Reddit users about A.I. One Reddit user asked about the potential roadblocks to Artificial General Intelligence, tech speak for A.I. that is as capable as human intelligence. John Smith, an IBM fellow and manager for A.I. tech, responded: “Deep learning is not a roadblock to Artificial General Intelligence (AGI), but it is not the answer. At this point in time, we don’t know how to achieve AGI or if or when it will be ever achieved.”
EYE ON A.I. HIRES
Postmates picked Ken Kocienda as principal software engineer of the online delivery startup’s X team, which is developing a robot that can make deliveries. Kocienda had worked at Apple for 18 years and recently wrote the book Creative Selection: Inside Apple’s Design Process During the Golden Age of Steve Jobs.
Cloudera appointed the data analytics firm’s chairman Martin Cole as interim CEO. Cole replaces Tom Reilly, who suddenly retired as CEO, a few months after the company finalized its merger with data specialist Hortonworks.
PA Consulting has hired Alex Vayner to become the head of the consulting firm’s Americas A.I. and automation team. Vayner was previously the chief of Capgemini’s Americas data science and artificial intelligence practice.
EYE ON A.I. RESEARCH
A.I. as a master chef. Researchers from Samsung and Rutgers University published a paper about using deep learning to create realistic looking, but fake photos of meals like salads and cookies, based on ingredients. The researchers’ A.I. system involves a popular A.I. technology called generative adversarial networks that involves two neural networks—software that learns from data—working in tandem to create synthetic images and videos.
A.I. is power hungry. Researchers at University of Massachusetts Amherst published a paper about potential energy concerns related to training deep learning systems to understand human language. The researchers “recommend a concerted effort by industry and academia to promote research of more computationally efficient algorithms, as well as hardware that requires less energy.”
FORTUNE ON A.I.
Microsoft Removes Face Recognition Photos Amid Privacy Controversy – By Jeff John Roberts
Walmart Set to Deliver Groceries Into Your Home – By Phil Wahba
Google Is Changing Its Search Results. Here’s What to Expect – By Alyssa Newcomb
Amazon’s giant Sudoku puzzle. Wired visited one of Amazon’s cutting-edge warehouses in Colorado, where robots and people work together to ensure that customers get their packages delivered on time. To help the robots move in the warehouse, the company created a cloud-based coordination system that directs where each robot goes, when they should recharge, and when they should get back to work. An Amazon robotics manager describes managing the robots as “basically a very large sudoku puzzle.”