On November 6, the “2018 AI + New Media Forum” hosted by The Fourth Paradigm was held in Beijing. At the conference, the fourth Paradigm released the AI product of enabling media — Recommendation Intelligent Recommendation System (hereinafter referred to as recommendation). The product integrates advanced artificial intelligence algorithms to help media quickly and accurately match users’ reading needs, improve media operation efficiency and user retention time, realize user increment and traffic realization, and accelerate media’s progress toward “intelligence”.

Insight into the real needs of the heart AI recommendation system has become the basic ability of media

With the opening of the era of mobile Internet and the intensification of competition in the media industry, problems such as the serious homogeneity of the constantly emerging reading client content and the increasingly fragmented browsing time of users have become prominent. How to retain users and meet their personalized needs for high-quality content has become the key to operation. Previously, the industry generally regarded technology as an important productivity to improve content and user operation efficiency, and the recommendation system based on artificial intelligence and other emerging technologies has become a sharp tool for the “intelligent” transformation of media. Some insiders point out that the algorithm based on artificial intelligence is no longer the core competitiveness of news clients, but a basic ability.

In fact, intelligent recommendation of “thousands of faces” is one of the important breakthroughs in the game between user attention scarcity and information overload. It has become the basic development logic of new media that “constantly optimizing iterative AI algorithms — increasing user interaction behaviors — obtaining richer data — pushing more accurate data — more inseparable from users”. However, due to the high threshold of ARTIFICIAL intelligence technology and the lack of relevant talents, it is difficult for the media to build their own technical teams to build recommendation systems. The recommendation released this time provides a good platform.

Commend to upload content, content management, content distribution, recommend intervention, front-end rendering recommended one-stop service visualization platform, support for PC, WAP, APP full access to the platform, help the media from 0 to 1 set up recommendation system, significantly enhance the user active, retained and watch how important business indicators, while reducing costs, Greatly improve the efficiency of media operation, so as to realize the transformation of business intelligence. At present, Xianjian has carried out in-depth cooperation with Xinhua News Agency, People’s Daily, Huanqiu.com, Ti Media, Huxun, Yiou, CSDN and other media. Among them, in CSDN information flow scenarios and WAP site advertising recommendation projects, the first recommendation achieved a 110% increase in click rate, while the page view and revenue increased by 187% and 49% respectively.

Low threshold, advanced algorithm, multi-scene application recommendation to enhance media “hard power”

In order to solve the high threshold of artificial intelligence application and development cost big difficult problem, will be the fourth paradigm “prophet” in the fourth paradigm ability of data processing, image processing, natural language processing (NLP), large-scale machine learning framework, deep learning, automatic machine learning (AutoML), automatic tuning and advanced algorithms such as encapsulated in the first recommendation, by the user interface, It only needs three steps to complete the whole process of self-built recommendation system, eliminating the process of development and basic learning, and lowering the threshold of AI application in media. At the same time, Xianjian provides rich recommendation configuration and data statistics functions such as scene management, content management, rule configuration, online rendering, report statistics, etc. Users can know the recommendation status at any time and accurately control their own recommendation services.

It is worth mentioning that in order to avoid the occurrence of bad values such as pushing vulgar information, the Fourth Paradigm introduced the mainstream value algorithm, which was first introduced in the industry. Through the three steps of content quality control, intelligent distribution and communication control, it realizes the self-evolution of “people-oriented”. The algorithm achieves accurate matching between mass content and users’ personalized needs without removing the mainstream core value, making the algorithm more scientific, rational and controllable.

In addition, the recommendation support real-time and comprehensive analysis of user search habits, scene and context state, achieve accurate content distribution, PV (page views) increase up to 30-200%, user engagement and click depth significantly enhanced. Data updates at the millisecond level guarantee real-time recommendation of quality content. At the same time, the recommendation is also equipped with search engine optimization function, to achieve SEO intelligent optimization.

At present, Xianjian provides personalized recommendation, associated recommendation, popular recommendation and other three services. Based on the intention recognition capability of machine learning, recommendation can recommend the most suitable content for the current scene for each user at different times and places, truly providing personalized recommendation service of “thousands of fields and thousands of faces”. Based on semantic analysis of the current content, the associated recommendation recommends information similar to the current content to increase the stay time and the number of articles per person. Popular recommendations are ranked globally, by channel or by region based on the statistics of the number of views of the content on the site.

Xianjian is a one-stop intelligent recommendation product aimed at the real pain points of media. While exporting technology, it integrates industry insight and understanding, providing a solid guarantee for the “intelligent road” of media. As Fourth Paradigm founder and CEO Dai Wenyuan said, “The value of AI goes without saying. AI will play a huge role in the new ‘examination room’ of media, and recommendation will become a ‘conventional weapon’ in the media industry.”