(Baidu VR original manuscript, reproduced please indicate the source)
(Beijing time February 4, 2017) Recently, Xnor.ai, a Seattle-based artificial intelligence (AI) start-up company, announced that it has won a seed round of US$2.6 million in financing to develop the Xnor.ai platform with this investment fund. Using technology in areas such as drones, smart phones, and automobiles, this time, the team continued to integrate AI technology into various devices.
Xnor.ai startup AI technology
Xnor.ai’s investors in financing this time include Paul Allen’s Allen Institute for Artificial Intelligence (AI2) and Madrona Venture Group, of which Paul Allen is one of Microsoft’s founders, including the Institute for Artificial Intelligence (AI2). It is a new AI technology that Xnor.ai is co-developing and is currently seeking split independence.
About Xnor.ai This startup company is essentially a smart local computing mathematics. It can make AI-like visual and language models run almost anywhere. In the future, they want to make AI technology simpler and more mature. Simplifying mathematical operations to a rough equivalent in binary operations can improve the speed and efficiency of the AI ​​model by several orders of magnitude. And it has the potential to transform the industry.
Application of Xnor.ai technology on various devices
Ali Farhadi, the senior research manager of the AI2 project, demonstrated the results of their team's work. He took out a mobile phone to open an application and pointed it out of the window. At the same time, the mobile phone view was instantaneously covered with various external objects such as boxes and boats. , car, telephone, various tag information. But Farhadi explained: "In terms of memory, computing, and power, there is a big gap between the needs of AI algorithms and what other devices can do. Assuming that you use tags to tag pictures on your mobile phone, you need to send pictures to The GPU server is then processed and returned to you using tags. For this reason, Xnor.ai's technology has, to a certain degree, solved the ability and efficiency of this complex AI algorithm."
It is worth mentioning that Xnor.ai has not yet done a lot of accurate marking of physical objects on the outside world, such as the marking of front-segments, people and animals, but this is not a serious problem. Team consciousness Therefore, the CNN calculation does not have to be accurate. It should be that the result is a confidence level, not an exact value. The neural network that trains the identification ship does not reach the single value corresponding to BOAT, but the technology using Xnor.ai can completely imitate 80% of the outside world. Similar value, this is already very great.
Farhadi added: "Our goal is to make all the equipment around it smarter. It can be mobile phones, handheld devices, wearables, virtual reality, augmented reality, drones, cars and electronics. These things will be Driven by AI technology and applied to life."
Silicone Cable With Braid,high voltage cables,silicone rubber wire,High temperature shielding
Jiangyin City Weicheng Special Cable Co.,Ltd , https://www.weichengcable.com