Baidu updated AI strategy with AR cut UAV project
According to media reports, the recent deep learning of Baidu Baidu in Artificial Intelligence Lab Director Lin Yuanqing also outlined the main research directions: deep learning platform, basic image recognition technology, fine-grained image recognition, video analysis, image recognition, AR technology and hospitals.
Basic image recognition technology in Baidu’s OCR (optical character recognition) technology greatly improved, including mobile phone identification accuracy rate of more than 99%. In addition, Baidu would next month, thin-grained image recognition feature applied in glutinous rice: when user dish picture, near Baidu algorithm to recommend to users within 5 km and the restaurant.
Although Baidu wide deep learning laboratory in artificial intelligence direction, but there are some vulnerable project, which cut off the outside world look on the drone program. Lin Yuanqing believes that market potential for unmanned, but Baidu’s perplexity is how they compete with large boundary.
UAV positioning and it is closely related to the use of cut scenes, middle of last year, Baidu Union billion airlines launched drone takeout service users in support of unmanned delivery takeout store orders, delivery staff in the App, enter the destination of the flight, unmanned aerial vehicles will send out into the hands of users. Drone takeout from the test up to now has been a year of service time, outsiders said, judging from the present conditions, the services belong to false needs, and costs are too high and too many factors beyond their control, and so on. Baidu also imagine using unmanned aircraft mapping remote areas map, but after analysis found that the market value is not significant.
AR technology is Baidu’s plays. When it comes to AR, Lin Yuanqing replied, “it’s definitely can make money.” Currently this market and there were no large industry leader in Xinjiang, leaving the Baidu space is very big. Earlier this month, Baidu DuSee,DuSee released a mobile AR platform computer vision and depth of what is learning to understand real-time image, the platform can be via a Smartphone hardware to “understand” real-world 3D environment, which causes the phone to generate images with real-world interactions. It is reported that DuSee can be used in interactive advertising, marketing, education, manufacturing, areas such as interior design, future they AR and speech recognition, and natural language processing technology combining. Vera Bradley iPhone 5S case
Actions ranging from Baidu, whose force adhere to three principles in the field of artificial intelligence: AI technology used in existing products; cut off the market value of the project; avoid the competitive landscape has been set. Vera Bradley tech case
“DuSee is a natural extension of AI”, Baidu launched the AR platform DuSee
Interview with Baidu Wu Enda: what deep learning can bring to a company value
Originally published at ysl6.wordpress.com on August 29, 2016.