Greater computing speeds, data volumes provide fertile ground for deep learning
Deep learning is a relatively new branch of machine learning, but it will soon be an invisible part of every organization, according to a new report from the Future Today Institute, an organization that provides forecasts on how emerging technology will disrupt business and transform the workforce.
While the concept of deep learning isn’t new, what has changed recently is the amount of compute power and the volume of data that’s become available, said the 2018 Tech Trends Report. In practical terms, this means more and more human processes will be automated, including the writing of software, which computers will soon start to do themselves.
Deep learning has been hampered by the processing power of computer networks, the report said. As computers become faster—and as hardware architecture evolves—systems will perform tasks at super-human speeds.
"Some of the best use cases for deep learning now are found in complex data-driven projects,” said Amy Webb, founder of the Future Today Institute.
Deep learning is being used to mine the data within images.
“For example, Facebook and Google are both extracting a wealth of data from images that we post,” Webb said. “A stop sign isn’t just a stop sign; it’s red and white, it’s outside, it’s angled in a particular way."
Deep learning techniques will become more ubiquitous as they move from the lab into the mainstream over the coming years, Webb said. “I anticipate much of the early consumer and business-facing products to be within the realm of image, object and voice recognition, which will be used in security applications, information technology, real estate and banking," she said.
Because of the shortage in data scientists, companies should invest resources now into training staff and developing the next generation of workers, Webb said.
“Those who are hybrid-trained, who understand both data management and other specific areas of work, will be in high demand,” Webb said.