Technology

How An Aatmanirbhar Bharat Can Pick Up Pace In The Artificial Intelligence Race

Ravi Singh

Aug 31, 2020, 12:31 PM | Updated 12:42 PM IST


AI is the new frontier in the Make-in-India movement.
AI is the new frontier in the Make-in-India movement.
  • How India can plug the gaps in its artificial intelligence capabilities and leap ahead in the global race
  • After the clarion call for building a self-reliant India was made by Prime Minister Narendra Modi, one critical battleground has received scant coverage. The field of artificial intelligence (AI), with all its complexities, is a bit difficult for the media and, in many cases, think tanks to cover.

    The stories have swung from a dystopian vision of AI-powered job losses and futuristic visions at one end to some fledgling startups with not-so-impressive products at the other.

    Here, I will give, in simpler language, a strategic and tactical view of the gaps in India’s readiness for the AI race.

    Let's clarify the three terms that are a must for readers to understand the issue.

    Big data, defined as the power to store a large amount of data in data centres, which also provides the computing power to run systems and perform complex mathematical operations. This has accelerated AI development in the past decade.

    Machine learning (ML), a subset of AI that provides systems the ability to learn and improve from experience automatically.

    Deep learning, a subset of machine learning that uses neural networks to analyse different aspects with a structure similar to the human-brain neural system.

    Most of the latest research is in deep learning with a sprinkle of ML.

    In core research, I have seen paper after paper on computer vision and natural language processing (my core area of interest) and have rarely come across an Indian author.

    The research field, which although has its centre in the United States (US), has been increasingly dominated by researchers of Chinese origin (in China or abroad), with Canada and Germany following suit.

    In the past two decades, the increasing investments in research and development (R&D) by the Chinese government in universities has paid great dividends. The achievements of Chinese technology giants are rivaling those of their American counterparts.

    India, on the other hand, has some good universities but no AI research institution of major significance. No major AI model of note has an Indian researcher’s name.

    That being said, things are much better when it comes to applied AI. Indian startups have made some inroads and shown promise.

    The reason is easier availability and clear packaging of AI knowledge.

    • Freely available research papers and the corresponding code with knowledge sharing on Github

    • Open-source programming platforms by Google (Tensorflow), Facebook (Pytorch), and others

    • Cheaper and application-oriented courses, which were earlier restricted to the hallowed universities in the West, available on Coursera, Udemy, Udacity, and even Indian platforms like Upgrad

    • Easily available cloud-computing platforms

    Indian engineering students have started acquiring these skill sets. Despite the the ringing of alarm bells from information technology (IT) industry leaders a few years back, it seems the current and upcoming generations of IT technocrats are well prepared.

    The Indian IT industry will, by and large, maintain its dominance in the AI era from a services point of view. In product development, we occupy a much smaller share of the global pie.

    The National Education Policy 2020 has made some positive changes, especially the teaching of coding from Class VI onwards. But developing AI capabilities will require major changes on how education is prioritised at the university level.

    There are two objectives that the nation should focus on, each with a different policy direction.

    AI Research — Centres of Excellence

    AI research needs a combination of qualified and experienced professors, an army of researchers, monetary investments in internal computing centres, and, most importantly, a proven history of research excellence.

    Right now, the government has divided its education budget across universities with a varied degree of specialisation. There are small AI centres cropping up in many government colleges with budgets and talent spread thinly. The capabilities in these centres can vary and, in many cases, is limited. There is very little interest from major technology companies. This is not the right way to develop a global competitive edge.

    The government of India should choose, say, only three or four colleges or institutions to focus all its budgetary allocation on for AI research and shut down the rest. This restructuring is not easy and will face opposition from local stakeholders. But if India wants to develop indigenous capabilities, this will be an essential step.

    Such centres should have the following:

    • Monetary power to host renowned researchers around the year

    • State-of-the-art data and computing centres

    • Foreign exchange programmes

    • Competitive Masters & PhD global entrance exams

    • Consolidation of intellectual property

    • Placement cells to private and government institutions

    • Revolving door for private-sector individuals to get into specialised courses as students, researchers, or teachers

    • Research contracts from the private sector

    AI & Big Data Teaching in Engineering

    Technology companies have for long lamented the lack of availability of industry-ready students graduating from engineering colleges.

    Some colleges in India have already taken positive steps to partner with private-sector technology companies in inculcating integrated courses. However, this differs across institutes, batches of students, and is a bit of a hotch-potch.

    This lack of structure especially exacerbates the issues because the field of AI and big data is changing rapidly with major updates arriving every two to three years.

    Starting with all the government engineering colleges (which can be followed up by private institutions), the below structure could be adopted (assuming a year is divided into two semesters):

    • Industry-designed theoretical AI and big data courses to be completed by the third year (for all)

    • The theoretical course should be revisited every three years

    • The theoretical course should cover concepts and not necessarily the latest research

    • The end semesters could focus on applied AI and big data designed by two of the major technology recruiters for the institute

    • This course can be taken up by those who are interested in this field or are placed in such roles

    • Hub-and-spoke model connecting the research institutes mentioned above:

      • Training programme for engineering institute professors

      • A steady supply of the best talent from engineering colleges

      • Research institute projects with the best students across colleges

    Just like India built up its missile defence and space capabilities, AI is the new frontier for us in the Make-in-India movement. Ultimately, building up AI capability needs investment, direction, and commitment rather than just policy papers.

    Some of the measures suggested can be implemented sooner. This exercise should not be left entirely to individuals and the private sector, or else we will miss a great opportunity to establish India as the AI technology hub of the world.

    Ravi Singh has over seven years of diverse management and consulting experience with a focus on AI, big data, and analytics.


    Get Swarajya in your inbox.


    Magazine


    image
    States