RAISE 2020: AI for Social Empowerment

In the history of human civilization, Artificial Intelligence is considered to be the next giant technological leap similar to electricity and the internet. AI is slowly possessing more power to radically transform the economic and social fabric of the world we live in.

The Responsible AI for Social Empowerment (RAISE 2020) virtual summit was a Global Artificial Intelligence summit organized by the Government of India in partnership with Industry & Academia. It was inaugurated by the Hon’ble Prime Minister Shri Narendra Modi on October 5, 2020 at 7 PM. The summit was a global meeting of minds to exchange ideas and charter a course to use AI for social empowerment, inclusion, and transformation in key areas like Healthcare, Agriculture, Education and Smart Mobility amongst other sectors.

It witnessed robust participation from global industry leaders, key opinion makers, Government representatives, and academia. The RAISE 2020 Summit featured some of the most exciting start-ups working in Artificial Intelligence and related fields.

C-DAC – A Make In India force that’s powering India’s AI revolution

Sindhuja Balaji, a Senior Content Writer with India AI, writes:

The Centre for Development of Advanced Computing (C-DAC) is intent on setting the benchmark as the facilitator and manufacturer of indigenous high performance computing capabilities, and harnessing its decades-worth experience in Indian language computing – powering a homegrown AI revolution

32 years ago, The Centre for Development of Advanced Computing (C-DAC) was set up with the primary goal of building supercomputers. Over the years, C-DAC has set out on many initiatives including the Indian Language Computing Solutions, health informatics and applications of technology & electronics.

Today, C-DAC is pioneering R&D in the information technology and electronics domain in India. At RAISE 2020, C-DAC Director General Dr Hemant Darbari spoke at the MeitY showcase session, on the various accomplishments of C-DAC, and revealing its true might as the entity that could power the AI revolution in India.

Param Siddhi AI: C-DAC will be commissioning India’s largest HPC-AI supercomputer ‘PARAM Siddhi AI, a machine with 210 AI Petaflops (6.5 Petaflops Peak DP).

Param Siddhi AI will be based on the NVIDIA DGX SuperPOD reference architecture comprising of 42 NVIDIA DGX A100 systems, connected with NVIDIA Mellanox HDR InfiniBand networking along with indigenously developed HPC-AI engine, Software Frameworks, Cloud Platform by C-DAC. Incidentally, C-DAC started its foray into supercomputing with PARAM 1GF in 1988.

Param Shavak: The Param Shavak machine aims to provide computational resource with advanced technologies to perform high-end computations for scientific, engineering and academic programs that can address and catalyse research using modelling, simulation and data analysis.

This initiative is expected to create HPC-aware skilled workforce and promote research by integrating leading-edge emerging technologies at grassroot level. This has 317 Teraflops peak computing power with 8GB storage, and powered with 2 multicore CPUs each with a minimum of 14 cores. There are four variants of Param Shavak – HPC System, DL GPU and VR system (virtual reality). Param Shavak Srishti – a gen next bio informatics appliance for bio and agriculture was launched by MoS Sanjay Dhotre in Jan 2020 in Maharashtra.

National Supercomputing Mission (NSM) HPC Infrastructure: This includes the Param Shivay 650 Petaflops supercomputing facility at IIT Varanasi and another 650 Petaflops Param Brahma Supercomputing facility at IISER Pune. A 1.3 Petaflops Param Shakti Supercomputing facility at IIT Kharagpur. So far, more than 1800 supercomputing faculties have been trained and cumulatively trained over 573,000 HPC jobs.

Exa-Scale Computing Mission: C-DAC is on a mission to establish dependable and secure exa-scale ecosystem designs and disruptive technologies– this includes the design and manufacture of server boards, interconnects and storage with embedded silicon photonics to make India a leader in high end software development.

Build Approach: Now Phase 2 of these systems are being manufactured in India, and Phase 3 systems are being designed and manufactured in India. The Rudra server, developed with the Intel processor and pilot HPC system was launched last month, and Phase 3 systems are being planned more than 30 Petaflops. In addition, the indigenous products include Trinetra HPC Interconnect, HPC software stack development and HPC processor.

AI and Language Computing: One of the earliest projects spearheaded by C-DAC was Indian language computing. With superior advancements in hardware and software today, including AI and NLP, C-DAC is keen to leverage this experience over thirty years to develop Bahubhakshak Speech to Speech Translation Systems and IE/IR systems for agriculture, healthcare, education, legal and banking domains. In addition, C-DAC wants to facilitate and mentor start-ups to develop AI based solutions for various use cases that can cater to India’s diverse population.

HPC for Social Good

The applications of C-DAC's HPC are vast, and have a significant impact on social sectors. So far, C-DAC has been able to apply HPC capabilities to develop flood warning and prediction systems for river basins in India, seismic imaging to aid oil and gas exploration, urban modelling, genomics and drug discovery, materials and computational chemistry and the design and development of Direct Contact Liquid Cooling systems.

Education and awareness for Responsible AI

Samiksha Mehra writes;

Day two of the five-day virtual summit on AI- ‘RAISE 2020’ saw some insightful discussions between the best AI minds in India and the world. One such panel titled ‘Education and Awareness for Responsible AI’ highlighted the key concerns in AI adoption and the potential models of awareness generation across the public sector, private sector, academic institutions and the general public.

Hosted by Mr. Jaideep Mishra, Joint Secretary & Group-Coordinator at Ministry of Electronics & Information Technology, the session began with a keynote speech by Mr. Rahul Sharma, President of Amazon Internet Services Public Sector for India and South Asia. The illustrious panel, moderated by Ms. Urvashi Aneja, Founding Director of Tandem Research, consisted of Dr. Rahul Panicker, Chief Research and Innovation Officer at Wadhwani AI, Dr. Rohini Srivathsa, CTO at Microsoft India, Mr. Kye Anderson, Strategist for Major Impact Initiatives at AI Sweden.

Giving a historical perspective of the evolution of AI and ML through Amazon’s 20-year long journey with these technologies, Rahul Sharma in his keynote address noted that “there isn’t a single business function at Amazon that hasn’t been made better through machine learning.” He highlighted the four key elements for enabling wide adoption of responsible AI: Breeding a culture of innovation empowered by Cloud technologies; identifying the right use cases and business problems, and nurturing start-ups; making available ML-ready open data, and spearheading research and skilling initiatives for AI.

The panel discussion started with opening remarks from Dr. Rohini Srivathsa who spoke at length about Microsoft’s efforts to keep in check the monumental implications that a technology ,as disruptive as AI, is capable of posing. In addition to identifying six principles for responsible AI as Fairness, Reliability & Safety, Privacy and Security, Inclusiveness, Transparency, and Accountability, Microsoft has, under Satya Nadella’s direct leadership, chartered a course to bring these principles into practice. She hailed the policy initiatives such as WEF’s Global AI Council and NITI’s Responsible AI for ALL working document and called for societal, governmental, academic and international engagement to agree on the question of “what computers should do, not what they can do.”

Dr. Rahul Panicker outlined two aspects of responsible AI in his presentation:

Firstly, by positing responsible AI as a behavioural change comparable to wearing seatbelts, he emphasised the need for rules, tools, and norms, explaining each element with examples from the real world.

Secondly, he outlined a four-phase evaluation process for AI solutions, the most crucial being the final one, i.e. post-deployment monitoring (such as feedback loops) which mandates the technologists to “take responsibility for the consequences of their creation.”

Kye Anderson brought in a Swedish and EU perspective to the session with his brief presentation on the roles and responsibilities of the Swedish National AI Centre. Terming AI as the second big technological change since the advent of the internet, he hailed it as a societal challenge that needs more beyond merely research efforts in the form of cross-sectoral collaborations. He emphasised the need for proactive and futuristic policy, keeping ‘everyone’ including the future generations in mind.

As the session progressed into a robust discussion, the panellists answered the audience’s questions on topics ranging from AI regulation and the powers of the state to use of responsible AI in mental health and education sectors. The need to democratise the ongoing AI revolution by including those at the bottom of the pyramid in the design process was identified as paramount, as was the importance of developing and updating micro-skills and deep-skills for building core competencies. Urvashi Aneja gave her valuable insights on the algorithmic amplification and creation of echo chambers on social media and the subsequent need for holding these new-age media platforms accountable to the same standards of journalism as traditional media.

The panellists busted certain myths about AI and encouraged more and more people to experiment with AI for it has very little to do with just technology and more to do with the ethics and culture of its usage. The idea to leverage the power of storytelling and folklore emerged as useful for generating empathy and awareness around responsible use of AI. The audience was warned against the messages of polarisation, with Mr. Anderson adding that “those with the least have the most to gain from AI.” The panel concluded that going forward, we need to align our tech trajectories and AI regulations with societal and human values, and we need to do it fast because technological progress doesn’t wait for anyone.

Role of regulations for Responsible AI

Nibedita Saha writes;

Regulations are often viewed as causing obstruction to any innovation process. However, if the associated challenges are rightly and timely addressed, the innovation process can be accelerated. Given this background, the esteemed panellists, which had a balanced representation from the government, academia and industry, explored the potential approaches that can be taken in a bid to put in place proper regulations for responsible AI. The session was moderated by Dr. Arghya Sengupta, Founder and Research Director, Vidhi Legal who craftly steered the discussion to focus on the various regulatory approaches.

In his keynote speech, Mr. Brad Smith, President and Chief Legal Officer, Microsoft Corporation acknowledged the need for a national strategy based on a firm ethical principle. He stated that Microsoft has already been working on responsible AI and have identified 6 principles in the process. They are fairness and unbiasedness, security and safety, privacy, inclusiveness, transparency and accountability. Nevertheless, as AI is rapidly advancing, we are increasingly becoming AI-based enterprises facing multiple challenges. Such challenges need to be addressed by balancing the path between serving the people and following the fundamental value system.

While expressing his view on striking that regulatory balance, Mr. J Satyanarayana, Chief Advisor, C4IR India, World Economic Forum, mentioned 5 ways of doing so:

a) adopting a system of agile governance,

b) creating a developmentary sandbox (which will test an algorithm for its output),

c) establishing a set of ethical standard and principles,

d) creating an index that measures the balance and

e) being responsible by design.

Mr. Ralf Sauer, Deputy Head of Unit International Data Flow and Protection, laid emphasis on the fact that the way data is collected should have a human-centric approach as the concept of data protection is essentially based on the element of trust.

Dr. Gulshan Rai, who is the former National Cyber security coordinator, repeatedly highlighted the several cases of abuse of the AI technology, both in India as well as abroad. He reiterated the need to put in place an AI framework that should be modular as well as scalable. He also identified the need to create a uniform definition of AI as well as set up an independent body to test the unbiasedness of any AI algorithm.

Dr. Anand Rao, Global Artificial Intelligence Lead, PwC, focused on the trade-off between innovation and regulation and put forth a thought-provoking question by stating that are we using AI to displace humans or assist humans? Taking into consideration India’s demographic dividend, he voiced that India’s focus should be on assisted and augmented form of AI that has a human-centred approach.

Dr. Vineeth Balasubramanian, an Associate Professor, IIT Hyderabad brought a research flavour to the discussion by stating that current research is mostly being done to address the aspect of AI and a post-hoc and an ante-hoc models are mostly used in the process. He also added that since AI and ML are data-driven and lack human elements such as reasoning, common sense and background knowledge, it is important that these aspects are integrated in the process. He concluded by saying that there is a need to regulate not only the data but also the performance metric by which an AI is judged.

Conclusion

Shri Ravi Shankar Prasad, Union Minister, Electronics and Information Technology and Communications Minister was emphatic at the very outset, on why the theme was chosen after much deliberation.

In a country as vast and diverse as India, Responsible AI for Social Empowerment (RAISE) is critical, he said. The efforts have paid off big time what with 60k registrations from 139 countries. With this kind of a draw, we can expect a lot of action over the next four days.

Through DBT alone, the Government was able to transfer 172 billion USD to 370 million bank accounts that saved more than 23 billion USD, which would have otherwise been pocketed by middlemen. Such benefits have spanned across the economy and various sectors. The digitization efforts in the last few years have paid rich dividends, and we can see this in our lives every day – Aadhar, UPI, GSTN, and Digital Health Mission are some of the prominent examples. 10.6 million farmers have used e-marketplace to sell their farm produce. Widespread and continued efforts have brought in huge transparency, improved the quality of outcomes, and ensured a high degree of privacy.

While the pace of technology can be overwhelming, AI has immense potential to bring in economic development and in an equitable manner. Millions of lives are positively impacted, leading to inclusive development.

India’s greatest strength is Demographic Dividend, and the Indian IT Services sector has truly leveraged its potential. With the kind of skills we have today and the ecosystem support, we can lead in the AI space too. The ecosystem that we have built today is a testimony to the great progress being made in AI – CoEs, IPRs created; incubation centres, etc. augment the vision.

While corporate houses may look at these technologies to create solutions, serve their customers, and earn revenue, the Government wants to put a smile on the face of common people. A major step in that direction is when AI is used for better governance. A country of 1.3 billion people has a mobile user base of 1.2 billion, Aadhar Cardholders– 1.2 billion, and 600 million internet users. India has an abundance of data.

And it’s not just about large cities and towns. The development has to permeate our villages too. The Government is committed to connecting all villages through fibre optic networks in 1000 days which will be a massive step forward in building a robust digital infrastructure that will enable inclusivity.

The Private Data Protection Bill has been tabled and will soon be passed in the Parliament. This will be a major fillip to the digital economy- innovation can happen in a secure & safe manner.