The central government on Thursday launched two new platforms under its Indiai Mission: Indiai Datasetts Platform (Akosha) and Indiai Commute Portal.
Aikosha provides access to several datasets and models to develop artificial intelligence (AI) applications. Compute portal allows researchers, academics and startups to request access to powerful computing resources provided through more than 14,000 graphics processing units (GPUs) from 10 companies at a subsidized cost.
On the Aikosha portal, the Ministry designed an eight use cases cases to find out how datasets and models can be used to increase public services. For example, AI voice assistants for rural citizens can be made using datasets of voice samples from Indian languages and dialects. These assistants can help people use voice verification to apply for government schemes and prevent duplicate applications and fraud subsidies.
Also read Social Welfare Department to use AI for better implementation of public utility schemes
The AI Compute Portal, which was launched on Wednesday, will give researchers, academics and startups access to 14,517 Emphened GPUs. Till now, requests for access to computing resources were being sent to the Ministry of Electronics and Information Technology through email. These requests will now be processed through the portal, which will be completely operational within days.
Indiaai Mission was launched on 7 March 2024 with a budget 10,371.92 crores. Of this zodiac, 44% ( 4,563.36 crore) is determined to provide computing capacity through 10,000 GPU, which represents the largest of the seven columns of the mission.
“Teams have said that teams bidding for calculation capacity said they will add more GPUs every quarter,” said Minister Ashwini Vaishnav.
Also read How is Trump’s AI plan and China’s deep shaping Paris summit? Description
Yotta CEO and co-founder Sunil Gupta said that the company, which is providing 9,216 GPUs (about 63%of all approved GPUs in the first phase), will also apply in the second dialect round to close on April 30.
“Our progress for the creation of India’s Foundation model is moving very well. We have received 67 applications, “Vaishnav said. Twenty -two applications are for the development of large language models, while other small language models are for development. “An important part of computing capacity would be dedicated to develop our own foundation model,” he said, “he said, referring to large AI systems that power many applications.
Also read Government of Kerala to develop state -owned AI engine models for schools
IT Secretary K Krishnan said that unlike other countries, where most of the investment in AI was made by private companies, in India, the government has “invested in a clear, concentrated program, which is to implement the nationwide adoption of AI through seven dedicated columns.” He said that both the government and the private sector will continue to add general computing capacity “at cost, which is much lower than anywhere else in the world.”
The lowest bids for various GPUs were already more than 40% on average market rates. For researchers, students and startups, the government will give more subsidy to access to 40%.
Database
Currently, Aikosha hosts 315 datasets and 84 models from 12 outfits with several devices. Vaishnav said that the platform will have only non-individual data. These datasets include people from Digital India Bhashini Division, where a dataset consists of 1,684 hours of speech data in 12 Indian languages, which can help develop equipment for automated speech recognition.
Another dataset of the Telangana government includes transactions data from the state’s fair price shops.
Indiai CEO Abhishek Singh said Aikosha has been launched as MVP (minimum feasible product) with a limited number of dataset. “The forum exists, but it will need to be enriched,” he said. To develop applications on the basis of models and datasets, the division is collecting problems of problems from various ministries and state governments to present them as challenges as researchers, startups and challenges for the industry in the next round.
Also read Center eyes domestic AI model disrupts Deepsek Tech world as ‘cost-skilled’
To launch the platform, Matty preferred the dataset that researchers had requested in their applications under separate Indiai Mission columns, an officer on condition of anonymity to HT.
The ministry is not transferring all datasets from the current Open Government Data (OGD) repository, as Aikosha should have a ready dataset. “We are selecting what are relevant and A-Taiyar in terms of data quality, standardization and other parameters,” the official said.
Meity is training other ministries and state governments how to clean its dataset and use them to make applications for public services.
The model available on Aikosha includes several text-to-skills and translation models for Indian languages from Meity’s Bhashini Division.
Private companies like Saravam AI and Ola Krutam have also uploaded their models to Aastik. For example, Sarvam AI’s Shuka can be used to develop audio-based questions-answer-answer equipment in Indian languages. In addition to the Chitrartha English of Crutream, the text details in ten Indian languages can produce images.
To download the dataset, individual users must first sign up for Digilocker using their mobile number. Organizations will have to register through the locker, the unit of the mety using the Aadhaar. The purpose of this approach is to ensure platform security and dataset suitability. “It is a government platform. What if someone uploads a dataset with unfair images tomorrow? We need to be accountable, ”the officer explained. The platform allows contributors to set access controls for datasets and models uploaded by them.
On Aikosha, Meity has also underlined eight use cases, shown that various models and datasets on the platform can provide public services – such as to help understand the voice tributaries for rural citizens. This will require data on data and regional speech patterns of government schemes such as Dhashini and Project Wani (covering 54 Indian languages and dialects), e-ride and MNREGA.
Five teams were honored under India’s application development initiative for developing AI-based solutions for problems in five main areas: agriculture, climate change, health care, learning disability and governance. Teams with ideas on conceptual platform received 5 lakhs, while people with finished prototypes were honored 25 lakhs.
Efforts to create capacity
The Ministry also launched an AI qualification structure for public sector officials to build capacity among government employees about the life-cycle of AI development and implementation for public sector officials.
In addition, an AI component for the government’s online course platform for civil servants, Igot Karmayogi was added, which offers more than 1,900 online courses. Adil Zainulbhai, chairman of the government’s capacity building commission, said the new IGOT AI would use domesticly developed AI models to personalize the recommendations of the course for civil servants based on its ministry, role and title. The final goal is to make IGOT easy to use so that the authorities can only register questions, and the platform will suggest the most relevant courses.
“1.3 million civil servants have signed up for AI course, and 940,000 of them have completed a course on AI,” Zainulbhai said. He said that the Commission’s goal is to ensure that every civil servant in the country take at least one course on AI and “talk” in it by the end of the year.
Vaishnav also announced a call for applications for India’s Startups Global Acceleration Program, where, in collaboration with France’s Station F and HEC Paris, will spend four months in the selected startup station F’s accelerator program and get access to European investors and markets.
Principal Scientific Advisor Ajay Sood said AI faces three main challenges – skill development, computing capacity and quality dataset – which were addressed through Thursday’s announcements.
“India has been on the right path on AI and data,” he said. Accepting the confidentiality, security and moral concerns related to AI, Sood said that the final report on the AI regime will soon be released after reviewing all the comments submitted during the counseling period. He said that India would contribute the “technical-canonic structure” to the regulation of AI under the global data framework mentioned in the United Nations report, which is similar to measures recommended in India’s Digital Personal Data Protection Act.
