The academic legacy of AI in India
Vijay Chandru and Chiranjib Bhattacharya
Jujy 22nd 2019
The transformation of computer science from a discipline that began with the logical foundations of computing machines to a focus on the software needed to create applications that would leverage the increasing capabilities of these machines took place in the three decades following the early conceptualizations by Alan Turing during the war effort. Another transformation came in the mid-90s was when there was a realization that the internet was real, the transduction and digitization of all processes was inevitable and sciences had entered their fourth data-driven paradigm. The first three paradigms being empirical, theoretical and computational. Data Science rapidly became the dominant force within computer science.
“Computer Science has not been about computers since the 1990s”
Professor Christos Papadimitriou, Columbia University (IISc, January 2018)
Data Science has become the academic euphemism for the popular nomenclature of big data, analytics and more recently even AI as bandied about by marketing professionals, media and social commentators. In India, the premier academic conference that deals with these topics is called CODS-COMAD (Data Science and Management of Data) and is organized by ACM India. The 2020 version is in Hyderabad in January.
The world has been taken by storm by the power of AI with the amazing hardware advances that have made deep learning viable today. The successes in image recognition (Deep face®), gaming (deep blue ® and alpha go®), autonomous vehicles (Tesla®), robotics (da Vinci®), voice interfaces (SIRI® and ALEXA®) and natural language processing engines (Watson®) have been truly impressive and there is hardly a domain today that has not considered or is actively considering its “AI Options and Strategy”.
The idea of using observation, learning and computation to solve hard problems rather than causative theories based on laws and axioms and a formal deductive logic to arrive at predictions and solutions, is an Indian legacy that has been well documented. Historians of classical Indian science have called this computational positivism and it is well illustrated in its roots in Indian astronomy. Recent approaches towards AI and Technology can be seen as a massive scale revival and de-skilling of this methodology through exponential advances in computational machinery and algorithms.
We have chosen examples below of how this classical legacy has played out at the Indian Institute of Science (IISc). We could just as easily have used the Tata Institute of Fundamental Research (TIFR) in Mumbai to illustrate the academic legacy that sprung from Homi Bhabha’s prescience in initiating the TIFRAC project that led to the first functional state of the art computer (automatic calculator) built in India in the early 1960s. However, it was the department of Computer Science and Automation at the IISc established in 1969, just about when ARPANET was getting established, that would play an important role in nurturing the development of data science and artificial intelligence in India. The story told here has local versions at several other institutes of eminence such as the IITs around the country as well.
·Early foundational work in intelligent systems, artificial intelligence, machine intelligence and machine learning in India was carried out from the late 60s through the early part of this millennium in computer science and systems science at the departments of Computer Science & Automation as well as Electrical Engineering at the Indian Institute of Science.
o First course in the country on AI in 1970's by Prof G. Krishna.
o First course on Pattern Recognition by Prof. B.L. Deekshatulu in early 1970s.
o First course in Machine Learning in late 1990 by Prof PS Sastry.
o Foundational work on Learning Automata (precursor to reinforcement learning) by Prof. MAL Thathacher. Extended to networks of learning automata with Prof PS Sastry.
o Over 15000 citations of paper by Prof M Narasimhamurthy on Clustering methods (a foundational technology in AI).
o Machine Learning on Big Data: Several algorithms designed from IISc. Most noted contribution was by Prof SathiyaKeerthi along with his students who designed algorithms for learning SVMs (support vector machines) from large datasets. The algorithms are incorporated in several active software platforms including LIBLINEAR
o Textbooks by faculty on learning automata (Thathachar, Narendra, Sastry) automated theorem proving (Chandru & Hooker), machine learning (PS Sastry), automated manufacturing systems (Viswanadham, Narahari) and market mechanisms in game theory (Narahari).
o In Machine Learning and Data Mining, IISc comes within top 10 (9th) universities in Asia (76th in the world) while IIT B comes close second at 14. A recent alumna of the ML program at CSA in IISc, Himabindu Lakkaraju, just made the MIT TR 35 under 35 list for her work in AI and bias.
o World class research was done in computational linguistics, machine translation, natural language processing at MCC in Chennai, IIT Kanpur and later IIIT Hyderabad starting in the 1970s.
o Between IISc, IIT-B, IIT-M and Strand, India has repeatedly fielded winning teams in the global competitions for data mining – the KDD Cup as well as the SIAM data mining conferences.
The Department of Computer Science and Automation (CSA) at the Indian Institute of Science in Bangalore is about to celebrate its 50th (golden) anniversary and we felt that this is a good time to remind ourselves, entrepreneurs and investors, science administrators in the country and society at large that there is a deep intellectual foundation in these topics and capacity to innovate that can be drawn from. It is important to approach intelligent machines with broader strokes that include AI (artificial intelligence), IA (intelligence augmentation) and II (intelligence infrastructure or cyber physical systems) since all of them matter for the future of commerce, social well-being and scientific advance.
Experts will confide that India seems to have fallen behind the global leaders in the cognitive era of Industry 4.0. Our IT industry has had an enviable position of being the custodians of the world for back end data processing and yet we did not build an innovation engine like China has. The question facing us is whether we can compete in this race or will we end up becoming the scalable “Mechanical Turks” of data annotation and get paid at the lowest end of the value chain in this technology revolution as well.
Exactly a year back and with its usual panache, the leadership of NITI Aayog launched a unique brand #AIforAll premised on the proposition that India, given its strengths and characteristics, has the potential to position itself among leaders on the global AI map. The focus on five sectors that are envisioned to benefit the most from AI in solving societal needs: Healthcare, Agriculture, Education, Smart Cities & Infrastructure, and Smart Mobility & Transportation.
The report entitled “National Strategy for Artificial Intelligence” does acknowledge that “India has the necessary building blocks to develop a thriving AI research and development ecosystem, viz. availability of highly educated talent pool, world class educational institutes and an illustrious list of top notch IT companies dominating the global IT landscape. Despite these advantages India sees itself lagging considerably in producing world-class research and innovation in most technology fields, more so in AI.”
The NITI report goes on to propose a new structure based on centres for transformational AI, managed by industry at the commercial end, with feeders of innovation being centres of research excellence based in the academic institutes of eminence. The real secret of getting this translational and transformational impact of technology innovation appears to require academic researchers who are willing to step out of their comfort zones and jump in with both feet to engage in genuine partnership with industry in this flow of innovation from research to commercialization.
As it happens, Strand Life Sciences was a spin off from the CSA department of IISc in 2000 by four faculty members who translated cutting edge research in data science and AI, leveraged machine learning and natural language processing and impacted the life sciences and healthcare. The AI stack Strand Ramanujan® drives the revolution today from population to precision medicine in India thereby directly addressing one of the challenges of #AIforAll in Healthcare: increased access and affordability of quality healthcare.
In speaking in Bangalore on techno-entrepreneurship some years back, Dr. R. A. Mashelkar had this to say: “We have two goddesses. The Goddesses of Knowledge and Wealth. Normally they are treated separately and kept together. But there is a route from the Goddess of Knowledge to the Goddess of Wealth. That was how Silicon Valley was created. Somehow, we have not understood that. In India, we have not got the dynamics of techno-entrepreneurship. There are a few examples like Strand Life Sciences or Bangalore Genei or Tenet. What we need to focus on is how do we promote that and how do you make it a rule. This is our biggest challenge.”
About the Authors: Prof Vijay Chandru is INAE Distinguished Technologist at IISc and a co-Founder of Strand Life Sciences, India’s precision medicine solutions company. Prof Chiranjib Bhattacharya is Professor of Computer Science & Automation and co-Chairman of the committee that tabled the Karnataka Action Plan for Machine Intelligence. The views expressed here are personal. firstname.lastname@example.org email@example.com
 Narasimha, R., ‘Axiomatism and computational positivism: Two mathematical cultures in pursuit of exact science’, Economic and Political Weekly, 2003, 3650–3656  https://parsikhabar.net/science/dr-homi-bhabha-let-there-be-computing/3164/  http://csrankings.org/#/fromyear/1970/toyear/2019/index?mlmining&world