Yi Tu Network Technology Co., Ltd. (hereinafter referred to as "Yi Tu") recently announced that Yi Tu Science and Technology has completed the financing of $654.38 billion from Xingye Guo Xin Asset Management Company. Before Xingye Guo Xin invested in asset management, Yi Tu also received a $200 million C+ round of financing in June, with investors including Gaocheng Capital, ICBC International and Puyin International.
Yi Tu was founded on 20 12. It was co-founded by Leo Zhu, Ph.D. in statistics at UCLA, postdoctoral fellow at MIT's Artificial Intelligence Laboratory, and Lin Chenxi, former technical director of Alibaba Cloud. As shown in the figure, artificial intelligence companies with the most comprehensive technical layout at present include computer vision, natural language understanding, speech recognition and artificial intelligence chips.
2065438+May 2007, according to the figure, we obtained 380 million yuan of Series C financing from Gao Yan Capital, Yunfeng Fund, Sequoia Capital, Gao Rong Capital and Zhenge Fund. Among them, Zhenge Fund also participated in the angel round financing of several million yuan on 20 13.
Within 35 days, billions of dollars in financing have been completed. What is the ability of this artificial intelligence benchmarking enterprise rooted in Shanghai? As a technology company, why did you choose Shanghai as the base for your own business? How did Shanghai help the artificial intelligence business?
For Yi Tu, the recent good news is not just the smooth financing. In June this year, in the latest test results of face recognition algorithm, Yi Tu won the face recognition championship, which is the second time that the company won the face recognition championship in NIST competition. In the test results officially released by NIST from June 2065438 to June 2007, the recognition accuracy reached 95.5% under the false alarm rate of one in ten million, which was the best level in the global industry at that time.
The National Institute of Standards and Technology (NIST) is directly under the US Department of Commerce. Its main task is to establish national measurement benchmarks and standards, provide testing technology for American industry and national defense, and participate in the standardization technical Committee to formulate standards.
In the face recognition algorithm test under the guidance of NIST, all the data come from real business scenarios, which means that the test results represent the performance of the technology in actual combat scenarios; The data scale is achieved by sampling tens of billions of samples, reaching the order of millions. Because of its rigor, consistency and comprehensiveness, this test has become the largest, strictest, most competitive and authoritative face recognition algorithm competition in the world.
According to official reports, this year, according to the map, this indicator has been raised to a level close to the limit, that is, in the case of one million false positives, the recognition accuracy rate is close to 99%. It is also worth noting that the algorithm that won the championship according to the map last year is still ranked in the top ten in the test results in June this year after a lapse of one year.
Source: Phoenix Net