Transforming agriculture with big data

Increasing farm productivity is the key to meeting the food and energy requirements of the world. At IBM Research India, we are using data analytics coupled with IoT and AI to generate new and deeper insights for farmers and agribusinesses throughout the supply chain.

IBM Research and the Watson Decision Platform for Agriculture

IBM Researchers Receive Aegis Graham Bell Award for Work in Digital Agriculture

Learn more about IBM PAIRS

Transforming businesses with blockchain innovations

At IBM Research India, we are driving next generation data management techniques for blockchain platforms and solutions for international trade and supply chain collaboration. We are building decentralized and provenance-based platform for cross-organizational sharing of data and AI models. Our work covers high value application areas like supply chain financing, tax compliance and financial accounting processes.

Learn more about blockchain research at IBM Research

Towards educational transformation

At IBM Research India, our goal is to make learning deeply personalized and engaging place learner of all ages on successful pathways. Our work includes a dialog based intelligent tutor to support learning across domains and age-groups; vocabulary tutor for early-childhood-learning using speech and visual recognition based interactivity; Watson Enlight, at teacher tool which uses predictive analytics on longitudinal student data as well as a cloud play form to deliver highlight personalized career counseling and accelerate large-scale employment.

Transforming retail and distribution

From driving extreme personalization of the customer experience, to using blockchain in procurement, optimizing cataloguing with merchandizing processes, and using AI to support product design – we are enhancing and leveraging IBM’s cognitive capabilities to transform the retail industry.

Accelerating deep learning

With increasingly large training data sets and complex neural systems. Distributed deep learning at a massive scale is a critical need. At IBM Research India, we are working towards improving the scaling efficiency of IBM PowerAI systems and IBM Deep Learning as a Service cloud offering. With our latest results, we are able to train the resnet-t0 network using 256 GPUs in less than 45-mintues. This is the best known performance for 256 GPUs to date.

Fair and secure AI

How do we ensure that machine learning decision are fair and secure? At IBM Research India, we are creating tools for developers to investigate training data, transformations and models to detect, certify, and remove bias. We also provide tools for checking robustness of models against security threats like adversarial samples, data and model poisoning, as well as modelling theft.

AI for compliance

Can we teach machines to reason and perform regulatory tasks for compliance? We are working on non-parametric probalistic models for knowledge graph extraction, advanced neural models for text processing, reasoning by knowledge graph embedding and better search techniques, to build this artificial intelligence for compliance.

Automating business processes using AI

Business process transformation with the help of AI comprises capturing knowledge of processes, analysing, the process behaviour from this captured knowledge and re-designing the optimization points for process improvements. AI is immersive across three stages of (a) discovering process knowledge from structured and unstructured data like documents, emails , tickets and system logs (b) analysing the process knowledge to create actionable insights and (c) deploying the actionable via means like automation in order to create the new and improved process.