JD Zhizhi's Wise Open Platform: Innovating AI Data Annotation Services

JD Zhizhi, a subsidiary of JD Technology specializing in AI data annotation and collection, has developed the Wise Open Platform to address the labor-intensive nature of data annotation through technological and model innovations. The platform focuses on improving efficiency, reducing costs, and building trust between clients and annotation teams.

Technologically, Wise offers configurable annotation tools that cover various scenarios including 2D bounding boxes, pixel-level semantic segmentation, audio transcription, and text keyword annotation. These tools are currently free to use, with plans to expand the library to include features like image drawing, video annotation, and 3D point cloud annotation. The platform implements an efficient QAC (Quantized-Answer-Check or Quick-Answer-Check) workflow that enables real-time feedback between annotators and quality checkers, significantly improving annotation efficiency.

The platform also provides comprehensive online project management capabilities, allowing clients to publish projects, manage annotation and quality checking processes, conduct online acceptance, and handle settlements and exports.

In terms of platform model innovation, Wise addresses data security concerns through multiple strategies. The platform stores client data in isolated environments with financial-grade access restrictions, while annotation teams and individuals must complete enterprise/personal identity verification. Real-time risk monitoring automatically flags suspicious users.

To reduce intermediary costs, Wise connects clients directly with vetted annotation teams. The platform has selected 10 high-quality annotation teams with extensive experience and verified identities. Clients can communicate directly with these teams through online tools or register their own annotation teams as guilds to receive projects.

For payment security, Wise leverages JD Wallet Enterprise Edition to provide online settlement services. Project fees are frozen in client accounts upon project publication and transferred to annotation teams only after client acceptance, ensuring trust and protecting both parties' interests.

The platform also supports annotation teams by providing them with platform capabilities. Teams can configure tools independently, distribute tasks to their members, and benefit from the platform's rich tool library and efficient workflows. Currently, the platform does not charge service fees.

JD Zhizhi aims to become the most trusted AI data annotation platform, providing robust service guarantees to clients. The platform can be accessed via QR code scanning.