Recently, 4Paradigm officially signed a contract with People’s Daily to jointly create mainstream algorithms for new media. While ensuring the accurate matching of massive content with the individual needs of users, to realize the dissemination of high-quality content in mainstream media and promote the transformation and innovation of the media industry in the AI era. —
Dai Wenyuan, the Chairman of 4Paradigm, said that the media industry is not only about flow, but also needs to spread positive energy. Thus, we need to change the characteristics of the past recommendation algorithm that only optimizes clicks and optimizes user time, and adds elements of values to the algorithm. We are not only pleased to have the opportunity with the People's Daily to explore and meet individual needs, but also reflects the mainstream algorithm value judgments. While ensuring accurate matching of content and user needs, a correct balance is achieved between individual needs and group values.
Ding Wei, the director of the New Media Center of the People's Daily said that algorithms are prevailing in the intelligent era. From a thousand people see the same thing to a thousand people see their interests of unique thing, algorithms are reconstructing the logic and rules of information dissemination. We have cooperated with enterprises such as 4Paradigm to launch mainstream algorithms in version 7.0 of the People’s Daily client to promote the strategic transformation of the People’s Daily client from traditional media to smart media.
He emphasized the three characteristics of mainstream algorithms. The first one is the more quality content. The platform creators of the People's Daily are analyzed by uploading users to control the quality of the content from the source. In terms of content classification, it relies on the People's Daily new media team and review team to classify and identify content and establish a quality evaluation system. Meanwhile, with the help of artificial intelligence technologies such as semantic scene recognition and deep learning smart technology, problems such as content reshaping are solved. The second one is to understand your recommendations better, to perform multi-dimensional feature descriptions, and to achieve efficient and accurate matching of massive content and individual needs of users. Mainstream algorithms comprehensively characterize users' interests through long-term and short-term behavior changes analysis of users, and dynamically characterize users' current interests and preferences through real-time phenomenon estimation. At the same time, with the ability to enrich the user's offline mining interests label meet the needs of users diversified and personalized. The third one is a richer and more open information environment. Mainstream algorithms can provide users with a cross-collar knowledge system and break the barriers of information cocoon rooms. Through the establishment of a knowledge user system in the massive information, the integration of user behavior and semantic recognition, and the mining of causal relationships, to make the machine can form a stronger reasoning ability, so as to realize the recommended content, from point-to-point expansion to cross-domain presentation.
The 4Paradigm "first recommendation platform" through the real-time and high-dimensional operation of multiple systems such as the quality assessment system, the user-platform two-way interactive recommendation system, the text analysis system, and the user portrait system. A special recommendation system was built from 0 to 1 for the People's Daily, and it has been officially launched on the People's Daily news client. "First recommendation" is a recommendation system service platform based on large-scale machine learning produced by the 4Paradigm, which aims to lower the barriers for media to embrace new technologies. Currently, first recommendation platform has carried out in-depth cooperation with thousands of media and content platforms.
Artificial intelligence is rewriting the media format, and the form and communication mode of news products are being redefined. Content distribution affects the distribution of media traffic and benefits, and personalized recommendations have become the infrastructure of major media. The 4Paradigm and People's Daily’s innovative exploration in the field of algorithms further enhance the value of content dissemination.
Release ID: 89036455