Gasgoo Awards 2022 applicant: Annotator labelling platform from Magic Data

For the Gasgoo Awards 2022, Magic Data’s Annotator 6.0 has applied for the Top 100 Players of China's New Automotive Supply Chain.

Photo credit: Magic Data

Product:  Annotator labelling platform 

Description

Annotator intelligent platform is a new generation products based on artificial intelligence technology which R&D team of Magic Data has been working on for several years. It works as an efficient tool for customers in major industries to improve data processing.

The Annotator has four major highlights: 

Multi-Model Annotation labeling project management, visual management, intelligent labeling and quality check.

Multi-modal data labeling tasks can be configured, supporting audio, video-image labeling, label visualization, relationship dependencies, and multi-level labels, making multi-modal labeling more convenient and efficient.

It supports AI-assisted labeling of various data types, which greatly reduces the labor cost and time cost .

It supports quick deployment of the enterprise side, easy operation and maintenance during use, and AI-assisted efficient labeling, which can reduce the overall cost of enterprises by 50% and increase the efficiency by 100%.

Annotator® uses modular design and development, supports voice, text, image, and multi-mode data annotation, supports heterogeneous data management, and supports annotation production split in pipeline-based mode.

Photo credit: Magic Data

Customize the composition type of annotation data, and expand or access the corresponding annotation operation interface.

Provides a large number of customized label Settings for various data annotation requirements,

The modular architecture provides Restful interfaces and module interface standards, allowing users to customize annotation worktables that meet their own service requirements based on platform data standards.

Unique advantages

In terms of product services and technological innovation, the company has reached international standards in ensuring data security and obtained ISO27701:2019 certification at the beginning of 2021, which is the first batch of enterprises in the industry to obtain the certification.

Secondly, Magic Data puts forward the principle of 2-8 data which help enterprises reduce costs and increase efficiency, when training AI models, 80% generality and 20% customization should be considered in the training.

In terms of technical innovation, the Intelligent voice expert team has been deeply engaged in the industry for decades and released Annotator®5.0 intelligent annotation platform SaaS free version/privatized deployment version, which has been widely recognized by the industry , such as Ping An, Shell, and more than ten industry customers have carried out privatized deployment business. Thousands of AI trainers have started using the free SaaS version, which has been well received on the vendor side.

Application

Develop a common and unified data standard format, support the label system of heterogeneous data, and realize federated learning and other machine learning scenarios through standardized Restful interfaces after multi-parties cooperation.

Prospect

Based on the accumulation of various industries, support for various industry-specific annotation scenarios, accumulate available annotation tools, and adapt to the data annotation needs of most industries based on machine learning.