Intelligent search and push engine

Based on Baidu's technological advantages and traffic management capabilities, it provides multi scenario recommendation and search services integrating content information, goods and conversational recommendation for private domain traffic operation scenarios in industries such as finance, operators, the Internet, and media, effectively improving user activity, retention, and revenue.

POC version is 99 yuan/month. Come and experience the charm of recommendation engine try now

Use Now Cooperation consulting
Help Documents
  • Product Overview
  • Product Matrix
  • Application scenarios
  • Product Functions
  • Product advantages
  • Customer Stories
  • Cooperation consulting

Product architecture

 Based on user data, behavior data, and material data, the multi-party data source training three-stage "recall sorting integration" recommendation and search algorithm helps build a full link search and push system relying on Baidu's high-performance and reliable engineering architecture.
Product architecture

Product Matrix

  • Intelligent search and push engine

     Rely on the deep learning platform, improve the level of personalization through recommendation, content and search mode, provide developers with end-to-end RESTful APIs and SDKs, and help the business system improve the effectiveness of private domain traffic user activation, retention and transformation.

  • Interactive recommendation

     Based on Wenxin's one word model capability, it provides users with accurate recommendations and personalized question and answer services in the form of interactive dialogue, aiming to accelerate the transformation of commodity sales

  • Intelligent recommendation platform

     Based on the five services of data, recommendation, analysis, monitoring and model, we provide customers with an end-to-end one-stop recommendation service capability application and analysis platform, create a service closed-loop, and help business growth.

Application scenarios

Information flow recommendation
Interactive recommendation
E-commerce transformation
Content co construction

Scenario Overview

In omni channel content, combine user tags to quickly optimize content and accurately distribute the content expected by users, so as to improve the activity and stickiness of end users in the site.

We can provide

  • High adaptability of recommendation strategies, reducing enterprise development and trial and error costs
  • Accurately acquire users' minds and improve users' stickiness and activity
Information flow recommendation

Product Functions

  • Efficient access to data

    Self adaptive docking
    It supports flexible docking and real-time synchronization of user-defined data sources to ensure the quality of data use and quickly go online.
    Recommended material management
    Data management and application recommendation are supported by various businesses such as graphic information, equity goods and customized activities.

  • High quality recommendation model

    Model autonomous learning
    The model is self-learning based on incremental data every day, and the model automatically iterates to optimize self-learning to the optimal state.
    Flexible configuration algorithm
    Support recall algorithm without code arrangement, sorting algorithm and comprehensive display strategy.

  • Multi dimension effect optimization

    Adapt to a variety of tools
    The intelligent recommendation engine supports data analysis tools such as traffic, behavior and funnel.
    Improve operational efficiency
    Use data analysis to reach business indicators, adjust operational strategies and improve operational efficiency.

  • Robust system monitoring

    Stable monitoring capability
    Support to improve the efficiency of troubleshooting and solving abnormal problems through performance monitoring and business monitoring.
    Platform based operation and maintenance
    Form a platform based operation and maintenance, and ensure the stability of global monitoring services.

Product advantages

Baidu deep learning platform empowerment

Combined with the large-scale machine learning system of Baidu Flying Propeller, the customized optimization model of the recommended scene and operator empowerment, the performance is up to 10 times of the community version.

Multimodal content understanding

Large scale pre training model learning based on massive text, pictures, videos and other data to provide accurate content structuring ability, interpretable recommendation results and rule customization ability.

Platform for practical experience precipitation

With rich practical experience in finance, media, pan Internet and other industries, the platform flexibly supports the recommendation of all types of materials such as information, videos, and commodities.

Joint operation

With the introduction of information flow content materials, including more than 3.8 million content creators of 100 kinds of vertical content, covering various content forms such as graphics, text, video, etc., to promote the diversified development of mobile content and improve business income; Provide professional operation consultation, improve operation efficiency and support user insight analysis and refined operation.

Customer Stories

A domestic head mobile phone manufacturer

Baidu intelligent recommendation solution was adopted, and through sample construction, feature and model selection, effect optimization and online deployment, recommendations were made on the home page, details page and popular APP scenarios of the application market, successfully improving user click through rate.

A large joint-stock bank

Rely on the content+technology advantages of the intelligent recommendation engine to reconstruct the discovery channel in the bank's App. By introducing high-quality content, multi-mode fusion algorithm strategy, high-performance and high availability engineering architecture and joint operation, the user activity in the App is effectively improved.

A head radio and television group

Rely on Baidu NLP technology and search technology to establish a search platform to improve users' search satisfaction, viewing rate and subscription rate. At the same time, the in-depth learning model is introduced to optimize the recommendation system, carry out personalized recommendation for IPTV users, and further enhance the user's on-demand viewing and consumption stickiness.

Cooperation consulting

According to the requirements submitted by you, product experts will contact you to customize the exclusive product scheme.

Consult immediately