Artificial Intelligence (AI) has become one of those ubiquitous words of our society—technology majors tout it; banks highlight it; even washing machine companies flaunt it these days!
Whether it is a nudge from your email provider or a recommendation from your e-commerce engine or a service conversation with a chatbot or an automated diagnosis of the medical scan or an automated assessment of the risk profile of a borrower—AI already impacts us significantly in our everyday lives.
What’s next for AI? What are the next big applications, and of relevance to India? How does India fare in this AI race?
At itihaasa (www.itihaasa.com), we asked these questions to leading Indian AI/ML researchers and mapped the Indian AI/ML research landscape.
I see six areas of emerging research in AI in 2019.
1.Unsupervised Learning: AI/ML research requires lots of data which is curated and labelled. Humans do not need labelled data to learn. New research will focus on learning without supervision and creating never-ending learning agents.
2.Reinforcement Learning: A child learns by trial-and-error. New research around RL will use rewards and punishment as signals for positive and negative behaviour of the agent.
Another area of research is imitation learning, with humans in the loop, which makes it possible to teach agents complex tasks with no need for explicit programming.
3.Explainable AI: AI happens to be a black-box i.e. we do not know exactly how AI systems are arriving at their decisions. And their algorithms may be inadvertently influenced by human-biases and may thus adversely impact people’s lives, especially when it comes to applications like medical diagnosis or risk assessment for loan disbursement. Researchers in 2019 will increasingly focus on ‘explainability’ or ‘interpretability’ and study ethics and fairness of AI/ML.
4.Causal modelling of AI: The current statistical-mode of machine learning systems will combine with causal reasoning tools in defining the next paradigm of AI. AI systems need to develop common-sense!
5.Resource-efficient ML: Imagine a self-driving car. To decide whether to apply a brake or not, the intelligence in the car has to be local, not on the cloud. New research will develop algorithms that will make edge devices and IoT sensors smarter and address issues of bandwidth, latency, privacy, and battery power.
6.AI and Blockchain: Various business objects collect onto a blockchain and the data typically belongs to multiple owners. New research will focus on how we do secure, confidentiality preserving AI.
India is beginning to understand the power and importance of data. For example, a computer vision algorithm in the context of autonomous vehicles, working at about 80% efficiency in western conditions, may work only at about 40% efficiency on Indian roads. Our researchers are capturing data on Indian road conditions and are planning to use it for training autonomous vehicles.
Indian researchers are working on interesting projects that focus on Indian context. These include using AI for predicting the monsoon, modelling floods, and Indian language processing. Our researchers are also part of a project that aims to learn from the architecture of the brain and develop new AI specific computer architectures for improvements in processing speeds and energy consumed.
Our analysis of data from the Scimago Journal and Country Rank (SJR) for the period 2013 to 2017 shows that, India ranks third in the world in terms of number of citable documents in ‘Artificial Intelligence’ and ranks fifth in terms of citations. Our researchers are creating significant impact and I expect this trend to continue in 2019.
I expect significant efforts undertaken in 2019 to create India-specific AI tools, data-sets and regulations. Let the 2020s be India’s decade of AI.
Kris Gopalakrishnan is chairman, Itihaasa Research and Digital and co-founder of Infosys Ltd