Intelligent risk control empowers content security, the new Hunan "Smart Eyes" is officially launched!
2024-08-10
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
Hunan Daily, August 10 (All-media reporter Wang Mingjun, trainee reporter Luo Aimin, photographer Xu Xing) This afternoon, the "2024 Audio and Video Content Security Technology Seminar Annual Conference and New Hunan's 'Smart Eyes' Content Security Intelligent Risk Control Platform Launch Ceremony" hosted by Hunan Daily was held in the New Hunan Building. At the event, Fang Binxing, an academician of the Chinese Academy of Engineering, Liu Yalou, deputy director of the Provincial Party Committee's Cyberspace Administration, Yu Jinwang, deputy secretary of the Hunan Daily Party Group and editor-in-chief, Yang Jian, vice president of Tencent Group, Zhao Lei, deputy general manager of CCTV.com, and Ren Jiangchun, chairman of Hunan Wendun Information Technology Co., Ltd., jointly announced that the new Hunan "Smart Eyes" content security intelligent risk control platform was officially launched.
Network security is becoming the "chassis" of the national digital strategy, and content security is becoming the "steering wheel" of the national digital strategy. The new Hunan "Smart Eyes" content security intelligent risk control platform is an important scientific research result of the Hunan Daily Media Convergence Content Perception and Security Laboratory, Academician Fang Binxing's workstation, and the School of Computer Science of the National University of Defense Technology. The platform realizes the intelligent detection and full-process tracking management of sensitive information and political content expressions such as politics, military, violence and terrorism, contraband, and values in the integrated media content such as text, pictures, audio and video, providing full-scenario cross-modal human-machine collaborative AI solutions for managers, producers, and practitioners in the content ecology field.
Unlike traditional content detection systems based on keywords, the new Hunan "Smart Eyes" content security intelligent risk control platform takes the media vertical model as the core and adopts cutting-edge technologies such as natural language processing, machine deep learning, and intelligent semantic recognition. By building a cross-modal AI detection algorithm + independent review knowledge base, it realizes multiple special algorithm model calibration and multi-task parallel detection. It can accurately detect multimodal illegal content in all scenarios, solving the problem of time-consuming and labor-intensive manual review and difficult reuse of experience knowledge.
In addition, with the support of an intelligent semantic analysis engine, the new Hunan "Zhi Mou" content security intelligent risk control platform has developed an AI assistant that fits the risk control field by developing a risk control brain intelligent body. It can use the understanding and reasoning capabilities of large models to identify, evaluate and warn of potential risks, and continuously iterate the learning capabilities of large models through a closed-loop feedback system, providing users with diversified intelligent services such as knowledge base assistance, intelligent questioning, and risk warnings.
The construction and promotion of the new Hunan "Smart Eyes" has achieved remarkable results. The total number of platform detections has exceeded 66 million times, not only fully covering the Hunan Daily's own media matrix platform, but also providing audio and video content security detection services to multiple government and enterprise units in Hunan, such as the Hunan Provincial Fire Rescue Corps, the Hunan Provincial People's Procuratorate, the Hunan Education Television Station, the Xiangjiang Group, the Xiangshui Group, and 69 county-level media integration centers and other media organizations.
In the next step, the new Hunan "Smart Eyes" platform will continue to optimize model algorithms and risk control brain intelligent bodies in the political field, and actively explore industrial operations, aiming to provide content security and intelligent risk control full-process services to more media, government enterprises, universities and other institutions, and provide valuable references for regulatory authorities to improve the risk prevention capabilities of local enterprises and the level of content ecological governance.
