The development of digital technology and the ecological integration of all-media content. Audiences are immersed in a sea of complex information for longer, receiving information from different worlds. With the higher degree of freedom of expression of users in cyberspace, the mixture of true and false information has brought great challenges to social governance. In recent years, the Internet information department has continuously strengthened the governance of the network ecology, and a series of "clean network" actions have been launched, which has also put forward higher requirements for the development of in-depth content analysis technology. The Polar Chain "Divine Eye Public Opinion Analysis System" was born focusing on the content challenges of the Internet era.
While subverting the traditional communication pattern, it also ushered in a new growth point for "AI + content". Artificial intelligence has been deeply applied and practiced in many content fields such as content production, distribution, review, and human-computer interaction.
The core appeal of the birth of Shenyan is to directly attack the content security pain points, release the labor costs in the review process, provide more accurate and efficient AI content review services, and create a one-stop intelligent content security shield to escort the benign ecological development of the Internet.
Multi-dimensional AI technology based on computer vision deep learning algorithms, such as face recognition, natural language processing, OCR (optical character recognition), video structure parsing, patented keyframe detection, etc., multi-modal intelligent parsing of video content, capturing key data in videos through real-time AI tags, quickly identifying and detecting risky content in videos, pictures, and texts, purifying the network environment, and meeting the needs of different types of users to curb risks.
This core ability of public opinion content analysis is reflected in the results of different dimensions in the following two classic cases:
Comprehensively help mango TV content security system construction
As early as 2016, Jichain has provided Mango TV with diversified services such as new interactive advertising products "Lingxi", AI scene advertising, and new ip e-commerce model expansion.
As the only Internet video platform under Hunan Radio and Television, Mango TV needs to monitor the program content 360 degrees without dead angles on the one hand, and at the same time ensure that the information can be identified as soon as it appears.
It is based on this content review capability that Jichain uses rich video algorithms to customize and expand, and on this basis continues to build the Mango TV content security review platform. It can train face models locally and customize hierarchical review strategies for UGC and PGC, improving the real-time processing power of video content.
The platform fully complies with the various content security standards of Hunan Radio and Television, ensures the customized demand for the safe dissemination of Mango TV program content on the Internet, greatly reduces the intensity of content review, and increases the efficiency of review, and continues to promote the all-round improvement of the content security system in the field of Mango TV video.
Public opinion early warning deeply empowers the credibility of network information work
In the era of information index explosion, it is necessary to track the transmission path of public opinion information in real time and efficiently in order to quickly respond to network public emergencies. In the actual work promotion, the enterprise website forums in the jurisdiction of the single-region Cyberspace Administration have many forums and wide coverage, and the efficiency of relying solely on human monitoring and auditing is too low.
For a long time, Polar Chain has continued to pay attention to relevant rules and regulations, and has iteratively optimized content review technology with the times. In view of the content review pain points of a municipal-level Cyberspace Administration in East China, Jichain laid an intelligent network public opinion information collection network around the core technology advantages of Shenyan, and collected all-media content information on websites and social media public opinion platforms in the jurisdiction.
The network public opinion alarm and analysis service can comprehensively study and judge the recent public opinion according to the time period, type and frequency of the alarm, and use charts to intuitively display the development trend of public opinion, laying a solid data foundation for the future work focus of relevant departments. In addition to daily content moderation scenarios, you can also identify bad artists in the content in real time and accurately identify sensitive words and their common variants.
Mining public opinion in a point-by-point manner can more comprehensively display the in-depth information and potential risks of public opinion events, prejudge public opinion and public opinion, fully respond to the "clean net" call to action, and enhance the credibility and public satisfaction of regional internet information work.
Innovation drives the construction of a knowledge sharing circle for smart education
Public opinion monitoring and review is only one of the important modules in the field of content analysis, with the fierce evolution of digital transformation in various industries, Polar Chain continues to expand the application of AI technology to practice according to the new needs of industrial development. The emergence of "Internet + education" has effectively responded to the educational reform in the era of lifelong learning.
During the epidemic period, "distance learning" and "online learning" have put forward higher requirements for teachers' teaching, students' independent learning ability, and the landing of network technology in each region. As a result, Jichain Technology developed and launched the "Intelligent Recommendation and Q&A System", which was successfully applied to an online learning platform in Hunan.
This system can collect learner attributes, behavior habits, hobbies and preference data to generate user tags through a series of in-depth data mining, build accurate user group portraits, so as to personalize the recommendation of learning resources according to the needs of teachers and students, and intelligently answer questions in the process of user use, so as to truly improve the collaborative ability of teachers and students and the efficiency of online teaching.
In the future society, the relationship between education platforms and users is no longer a transaction relationship that provides products and services, but is closer to a value creation and sharing. In this project, Polar Chain builds a knowledge-sharing circle based on learner interests. Cluster course learners, intelligently recommend friends with similar interests and possible acquaintances, and invite learning groups with similar interests to be created or joined.
Jichain Technology hopes that through the construction of a "knowledge sharing circle layer", it can truly recommend influential learning resources for learners for a certain discipline or field, and realize the co-creation of knowledge value.
Continue to focus on content review to refine granularity
With the in-depth advancement of the ecological governance of network information content, the "clean net" action has been further strengthened. Based on deep learning, Shanghai Jilian Shenyan video content intelligent review system, Jilian further develops image recognition technology to improve the accuracy and efficiency of content review.
Jichain cooperated with an Internet platform in Hunan to provide content review and public opinion monitoring services for its Internet platform content. According to the unique AI algorithm, intelligent pornography recognition is carried out on pictures and videos, and the degree of pornography is scored, which is convenient for users to quickly process pornography-related content and fully reduce the risk of pornography. At the same time, it can also identify all kinds of violent terrorists and sensitive content of current politics, and provide customized solutions by establishing sensitive databases.
By detecting video content frame by frame and tracking the trajectory stream, on the basis of truly "reading" the video, it realizes low-cost, high-efficiency and high-precision video review and content capture. Use OCR identification technology to filter advertising content. Compare the picture with the massive sample map in the sample library to judge the image attributes.
In addition, around the pain points of customers in the actual business development process, Jichain has carried out a series of customized expansions in the field of content review: strengthen the management of key groups, focus monitoring by adding or importing user IDs, and automatically trigger early warning reminders when the speeches posted by such users involve the set content of concern. At the same time, it supports the early warning of users' psychological problems, and further refines the granularity of content review by monitoring the negative emotional content published by users and warning users of abnormal psychological trends.
The efficient, accurate, flexible and comprehensive public opinion monitoring system built by Polar Chain with Shenyan as the core escorts the benign and healthy development of the Internet platform and further enhances its market competitiveness.
Through the continuous innovation and development of Polar Chain AI technology, the audit accuracy of the Shenyan content intelligent review system is higher than 99.5%, which can achieve 360-degree review of pornographic content filtering, advertising filtering, text filtering, etc., and the efficiency is increased by dozens of times under the same accuracy.
In the Internet era of accelerated change, content analysis and review is the basic guarantee for the operation and development of Internet content. As a fast-growing AI unicorn enterprise in China, Jichain will be committed to the in-depth exploration of the field of artificial intelligence for a long time, relying on its strong innovation ability and cutting-edge insight into market demand, deeply empowering the construction of an open and innovative network ecosystem.