Cross-Language Big Data Engine

Cross-language Big Data Collection Capacity By integrating powerful global data acquisition, storage and cleansing capacity, the big data collection platform of GTCOM has accumulated tens of billions of data of multiple languages and various modalities, with sources covering news media, social platforms (Weibo, WeChat, Twitter, Facebook, etc.), e-commerce websites, forums, blogs in more than 60 languages from 200+ countries. The afore-mentioned platform employs the following technologies: multi-source heterogeneous data standardization technology, data source auto-discovery and categorization technology, distributed intelligent collection technology, semi-automatic extraction by extraction template technology and deep web collection technology, etc. These technologies are combined on the platform, which lay a solid foundation for the quick discovery, configuration, collection and cleansing of data source and result in hundreds of millions of data updated daily with second-level update interval. Big Data Analysis and Mining Capacity GTCOM has researched and developed multi-language segmentation rules, POS tagging, syntactic parsing, semantic role labeling, semantic disambiguation and so on; providing technical support for the cross-language text data mining with data processing algorithm model for various languages; carried out the NLP algorithm optimization targeted at statistics machine learning to promote the accuracy and efficiency of the understanding of natural language; researched the semantic search technology based on ontology to realize more intelligent information discovery; combined the machine translation, semantic disambiguation and other technologies to develop cross-language search algorithm and expand users’ information access range; utilized in an allround way the knowledge graph, statistics machine learning and neural networks to develop innovative technologies and products in various fields, thus enhances the system’s recognition level and independent service ability. Cross-language Algorithms The multilingual NLP algorithm platform adopts the frontier technologies in artificial intelligence and deep learning such as convolutional neural network, recurrent neural network, deep belief network, conditional random field, random forest and word vector model, in combination with hundreds of billions of global multilingual text corpus data resources, to build 11 accurate and highly efficient NLP algorithms that support the global TOP-10 languages and cover lexical, grammar, semantic and other levels, accurate and efficient. The multilingual NLP algorithm platform provides developers with efficient and stable algorithm cloud services to meet the needs of global information processing application at one-stop and provide solutions for cross-language information processing. DaaS platform of GTCOM GTCOM provides data services to third parties through the DAAS platform, including multi-lingual data classification, data management, tag addition and model analysis to support multiple types of data access such as domestic and international news, social contact, forum and blog, customize data service based on users’ requirements, conduct efficient structured analysis and processing of various types of data through the company's rich algorithm model (see the list below), support real-time updates and real- time processing of news, social contact and other high-aging data, and achieve second- level data push for customers. The platform supports various forms of data transfer protocols, which enable customers to customize their data needs. It adopts algorithm model management, data flow settings and workflow settings to establish an "information bridge" between enterprises, enterprises and users, thus achieving the information integration, exchange and sharing between different systems through the internal information-driven engine of the platform. It is featured by application-oriented, process-based and simplification. Attachment：Algorithm model list... SaaS platform of GTCOM Based on the data accumulation and analysis of the YeeSight big data platform and through enterprise media and public opinion monitoring, media analysis and commercial intelligence monitoring, the company has developed a series of industry applications, including news big data, public security big data, financial big data and enterprise big data, and provide users with SaaS account to enable them to quickly and easily achieve self-help data retrieval, analysis and other operations. The platform enables real-time coverage of news, Weibo, WeChat, forum, blog, Twitter and Facebook data, with its application system capable of querying the data of the latest 30 minutes and tracing back to the data 1 month before the enterprise’s ordering service. To meet the needs of negative data judgment and negative public opinion warning of enterprises, negative warning arrival can be realized through SMS, e-mail and other forms, while supporting multi-dimensional data analysis and subject analysis in terms of trends, geography, media, communication, social contact and comparison.