AIKosha IDP:

Marking a year of the IndiaAI Mission, Union IT minister Ashwini Vaishnaw on Thursday launched an all-in-one dataset platform AIKosha, among a slew of other initiatives aimed at improving access to AI. The dataset platform is one of the seven pillars of the Rs 10,370 crore IndiaAI Mission and aims to provide a unified portal for seamless access to datasets, tools, and AI models. Let's take a detailed look at the IndiaAI Datasets Platform (IDP).

What is AIKosha: IndiaAI Datasets Platform (IDP)?

In simple terms, the IDP is a repository where people can find a variety of datasets, models, and examples to support AI innovation. It also includes tools and tutorials in a development environment called an AI sandbox.

The Ministry of Electronics & IT describes AIKosha or IndiaAI Datasets Platform (IDP) as a platform that will provide computing, datasets, tools, and a secure sandbox environment to India's researchers, entrepreneurs, and startups, helping them build state-of-the-art AI applications and solutions.

This platform is a valuable resource for creators, researchers, and AI enthusiasts. It provides a wide range of datasets, each with an AI Readiness Score to help users determine if the data is suitable for their AI projects. This makes it easier for users to choose the right data for developing their models.

Additionally, the platform includes features such as easy content discovery, AI readiness scoring, secure access permissions, data encryption (both when stored and in transit), a secure API, and firewalls to filter out harmful traffic in real time.

Salient features of AIKosha:

The AIKosha currently hosts more than 300 datasets and 80 models. It aims to improve access to datasets and models from diverse sources and sectors to catalyse AI-based innovation and research, facilitate the development of AI solutions by sharing Toolkits, and highlight use cases to drive inspiration on applications that can be created using the resources on the platform. The salient features of the IDP are as follows:

Datasets Repository: Gain access to a repository of datasets with AI readiness scores across various domains for model development. Browse through datasets suited to different industries to support AI projects.

Model Hub: Discover, build, and publish AI models that can be reused, fine-tuned, and validated by others. Share your models to support AI development, improve usability, and drive continuous improvements.

Integrated Development Environment (IDE): Speed up AI development with an all-in-one workspace for coding, testing, and fine-tuning models.

Dataset and Model Management: Users can share and edit artefacts within a secure environment.

Scalable and Secure Platform: Enhanced with features like data encryption both at rest and in transit, ensuring seamless and secure data and model sharing.

Library of Use Cases: A resource to explore AI applications and solutions across different domains.

Collaborative Organisation Management: Users can create organisations for teamwork, with controlled access to datasets and models.

AIKosha: Roles and Rights:

The AIKosha platform can accommodate a diverse range of users, categorised into roles such as Explorers, Contributors, and Organisation Administrators, each with specific privileges.

#Explorers have view-only rights on the platform. They can browse and download artefacts (dataset, model, or other resources) but can’t make any modifications or contribute new content.

#Contributors have all the rights of the explorer in addition to permission to upload, modify, and manage artefacts on the platform. They can also approve or restrict access to the artefacts they contribute.

Usage rights for artefacts on the AIKosha platform are governed by specific permission settings. Artefacts can be categorised as Open, meaning they are discoverable and downloadable by all registered users; Restricted, where they are discoverable by all users but require explicit approval from the contributor for download; and Private, where they are only downloadable by the contributor.

#Organisation Administrators (Admins) have all the rights of a Contributor, plus they can approve/reject users' requests to join the organisation, manage artefacts published by the organisation members, and approve/reject publish and deletion requests for different artefacts. They can also promote/demote user roles within the organisation.

IndiaAI Datasets Platform: Registrations

To register as an organisation on AIKosha, one first needs to register as an Explorer. The organisation registration is then facilitated by Entity Locker. If the organisation is not registered with Entity Locker, an authorised representative must first sign up for the service, following which they can return to AIKosha and register as an organisation by raising a request on the Entity Locker's Explorer Dashboard.

The registration process for an organisation also involves a few additional steps. The "Register" button on the AIKosha website allows individuals to sign up with their phone numbers via an OTP. After providing the necessary profile details, the registration will be completed and they can browse the platform as an Explorer.

Read More :

Fear Of Work-Life Imbalance Holds Back Women From Pursuing Leadership Roles: Report