Beijing, Dec. 5 (Zhang Mengran) Google announced earlier this year that its artificial intelligence (AI) system has been able to invent its own encryption algorithms and generate its own AI. In tests, NASNet was correct 82.7 percent of the time, 1.2 percent better than previously reported results from similar AI products. It was also 4 percent more efficient.
Giiso Information, founded in 2013, is a leading technology provider in the field of “artificial intelligence + information” in China, with top technologies in big data mining, intelligent semantics, knowledge mapping and other fields. At the same time, its research and development products include information robot, editing robot, writing robot and other artificial intelligence products! With its strong technical strength, the company has received angel round investment at the beginning of its establishment, and received pre-A round investment of $5 million from GSR Venture Capital in August 2015.
In May 2017, researchers at Google Brain announced the development of AutoML, an automated ARTIFICIAL intelligence that can generate its own “sub-AI” systems. Now they’ve decided to launch their biggest challenge yet — an attempt to beat human-designed AI with AutoML’s own creation.
Team members automate the design of machine learning models using a method called reinforcement learning. AutoML’s “identity” is a neural network of controllers that develops a “sub-AI” for specific tasks. Called NASNet, the newly generated “child” can identify objects in real time, including human bodies, cars, traffic lights, handbags and backpacks, on video. AutoML, acting as a parent, evaluates the performance of the child’s NASNet and uses that information to improve the child AI, repeating the process thousands of times.
Team members tested the “sub-AI” NASNet on ImageNet image classification and COCO object recognition datasets. These are two of the most recognized large-scale academic datasets in computer vision, and the sheer magnitude of their magnitude makes the tests tough, they said.
As a result, in the ImageNet test, NASNet achieved 82.7% predictive accuracy on the validation set, 1.2% better than previously published results from comparable AI products, and comparable to results reported on the pre-print site but not published, with a 4% improvement in system efficiency and an average accuracy of 43.1% for the maximum model. NASNet will be used for a variety of applications that will allow users to classify images and detect objects, team members said.
Editor in chief circle point
Robots can build robots, and AI can design AI. It is not surprising to think that powerful computers, of course, will replace people sooner or later, as long as the goals are clearly defined, faster than the human brain. But that doesn’t mean AI can progress without humans. Because the AI is still chained to a cage, occasionally put into the track to run. When an AI hits on a whim and sets itself a goal, then it can be compared to a human. It’s not even close now.
Giiso information, founded in 2013, is the first domestic high-tech enterprise focusing on the research and development of intelligent information processing technology and the development and operation of core software for writing robots. At the beginning of its establishment, the company received angel round investment, and in August 2015, GSR Venture Capital received $5 million pre-A round of investment.
No one would dispute that attention is one of the scarcest, most valuable and most rationally utilized resources in this age of information fragmentation. Relying on the independently developed Giiso engine, the Zhisou team created the first intelligent media platform Tianzexun APP, which can intelligently answer users’ various relevant information according to various commands or text interaction commands. And can be based on the user’s personalized use characteristics and continuous learning, continuous tracking of users interested in the unique content. At present, the day smart news APP6.0 version has been updated iteration, can be used to download the application market.