Dr. Gautam Shroff, Vice President, Chief Scientist, and Head of Research at Tata Consultancy Services - Enterprise AI for Business 4.0: from Automation to Amplification

Abstract: Every industry today is both impacted by and also seeking exploit AI technologies already widely used in the ‘new economy’. Our “Business 4.0” framework guides this ‘digital transformation’ of traditional enterprises by focusing on extreme personalization, developing ecosystems and embracing risk to drive exponential value: Enterprise AI developed with agility and deployed on the cloud form the critical technical enablers for Business 4.0.

In this talk I shall describe in how TCS Research is applying AI in traditional enterprises spanning the spectrum from automation to amplification, beginning with our hands-on experience of developing and deploying a deep-learning based semantic system for virtual assistance as well as knowledge synthesis within TCS, at scale, on our internal collaboration platform. The underlying framework is also being used to develop conversational systems for a variety of scenarios in banking and insurance.

Natural-language interfaces are also changing the way data is accessed and analyzed, with natural-language interaction replacing dashboards and even programming. Our deep-learning based platform for intelligently reading from images, be they scanned documents, handwritten notes, photographs, or even engineering drawings, is also an example of program synthesis where rules are crafted in natural language using examples.

Next, I shall give some examples of how AI and IOT are transforming manufacturing, by using deep-learning for detecting, diagnosing and sometimes predicting faults, as well as how deep-reinforcement learning can be used to manage supply chains from source to shelf., across ecosystems. Finally, since risk and compliance forms a surprisingly large fraction of enterprise IT spend, we are using NLP and semantic modeling to automate large parts of this onerous exercise.

Last but not least, embracing risk is critical to deploying AI in the field, as I shall explain with some examples: it is this cultural transformation rather than technology deployment alone might be limiting the pace at with traditional enterprises ‘go digital’.

I’ll close with an overview of TCS Research, focusing on AI, and highlight some other research problems we are working on: From machines learning to negotiate contracts, or trade in inefficient markets with limited information; to how AI systems can be tested for safety and correctness. Going beyond IT alone, our work in life sciences includes discovering biological networks to aid in improved drug design or applying AI in the physical sciences to characterize manufacturing processes or even design new materials.