Diabetic retinopathy (DR) is one of the main reasons for avoidable blindness among adults in India, which has one of the world’s biggest diabetic populations. Early detection and treatment are critical to limiting the damage but a shortage of well-trained ophthalmologists, especially in rural areas, remains a challenge.
Artificial intelligence (AI) is coming in handy to bridge the gap. Sankara Eye Foundation India, a non-profit that aims to eliminate preventable and curable blindness, has collaborated with Singapore-based Leben Care to deploy a cloud-based AI software platform—Netra.AI.
Built on Intel-powered technologies, Netra, which means eyes in Hindi, can help identify the retinal condition in a short time with the accuracy level of human doctors, using deep learning. It can tell a healthy retina from an unhealthy one.
Not just Intel, several tech have turned to AI to help detect DR. Google helps healthcare workers detect diabetic retinopathy, with possibilities of AI algorithms to assist clinicians in identifying other diseases as well.
Microsoft Seeing AI research project is designed for no and low-vision community. Seeing AI uses computer vision, image and speech recognition, natural language processing and machine learning to describe a person’s surroundings, read the text and answer questions. It can also identify Indian currency notes.
These are just some of the examples of AI improving the quality of life in India. Not just in India, AI for social good is almost a movement across the world, aimed at providing quality education, welfare measures and healthcare while taking care of the environment.
Delivering the social goods
AI for social good is an increasingly popular theme. When Google Research India, an AI lab based in Bengaluru, was set up in 2019, the focus was to find ways to help build an artificial intelligence ecosystem.
AI has been a subject of immense interest not just for the technological breakthroughs but also for the wrong turns it could take.
Since its origins in the 1950s, AI has been viewed with suspicion. Will it make humanity redundant? Will humans become subservient to machines?
A lot of water has flown under the bridge since the 1950s and extensive research of the last four decades has led us to “AI for social good”.
Worries surrounding legal, ethical and safety issues forced the tech ecosystem to look for ways that AI could be shaped to benefit people and society at large.
Be it digital healthcare, advanced education, forecasting floods, wildlife conservation, securing marine life, or predicting wildfires, AI has a role to play.
Powering learning
Shalini Kapoor, IBM Fellow, IBM India Software Labs, says AI has penetrated every aspect of our lives. From transforming businesses to making a mark in societal impact, AI is leading from the front.
“In the field of education, we believe that AI can enhance learning environments by unlocking learning potential resulting in improved outcomes and better student experiences. IBM is working with several state governments and educational institutes to ensure AI is part of the curriculum to impart AI skills at an early age,” she says.
IBM has partnered with the Ministry of Education and NITI Aayog on an online initiative for higher education.
Aligned with the Skill India mission to provide last-mile connectivity for quality higher education, sâmShiksha, an AI-driven, personalised learning platform, is being designed for a holistic learning experience.
It has the potential to scale up to become a “single-point” source for higher education in India by serving as a virtual campus, Kapoor says.
“Along with CBSE, we have also developed a curriculum of artificial intelligence as an elective subject for Classes IX to XII. We have covered over 200 schools across 13 states and over 15,000 students have benefitted from the programme,” she told Moneycontrol.
The coronavirus pandemic has accelerated digital adoption. Rohini Srivathsa, National Technology Officer, Microsoft India, says India is on the cusp of a digital revolution.
“During the pandemic, we worked closely with several governments and public health authorities to enable citizen services with AI. The government of India’s Saathi chatbot and Punjab government’s COVA app that helped citizens with critical COVID care information are powered by Azure AI,” she says.
If you look at healthcare, Apollo’s 24×7 virtual healthcare platform uses Azure AI to offer last-mile healthcare service delivery across India, says Srivathsa.
Bridging the language divide
AI and data are at the heart of driving economic and societal inclusion for everyone, including the one billion people with disabilities around the world.
AI4Bharat, an AI startup in partnership with Microsoft, is building AI models for recognising Indian sign languages, creating one of the largest datasets on it. This project has the potential to integrate people with disabilities into the workforce, says Srivathsa.
Google is using AI to help out with languages. Many of us prefer consuming content and getting things done in our mother tongue but language continues to be one of the biggest barriers to access information on the web.
“Solving for languages, in particular Indian languages, represents many challenges,” says Partha Talukdar, a staff research scientist at Google Research India.
From the complexity and variation of scripts—from Urdu to Malayalam to Gujarati—to the sheer number of languages across the country, it poses an interesting challenge for AI and machine learning.
“It is also something we at Google have been focusing on for a long time, from launching Neural Machine Translation for 11 Indian languages back in 2017 to now giving users the option to access high-quality web pages originally written in other languages and see it in their preferred Indian language,” says Talukdar.
Scientists at Google Research India have come up with a new language model called Multilingual Representations for Indian Languages (MuRIL).
MuRIL helps in handling challenges like transliterations, spelling variations, mixed languages, and other usages often observed in the Indian context.
MuRIL supports 16 Indian languages as well as English. “We have also adapted MuRIL for better query understanding for spoken query understanding,” says Google’s Talukdar.
Srinivas Lingam, Intel’s VP, Datacenter & AI Group, says AI is improving English-language proficiency for some of India’s most needy children.
“A project we’ve been involved with is the development of ReadToMe, an AI-enabled software platform, which is supporting more than 25,000 Indian schools. Implemented for students from grades 1 through 12, classes using ReadToMe displayed a 20–40 percent improvement in English reading and comprehension skills,” he says.
Google’s Read Along, earlier known as Bolo, helps kids learn to read. First tested in India in 2018 and rolled out across the world, Read Along is an AI-enabled Android app. The virtual assistant in the app reads a story then listens to the child narrating it back.
Using AI-based speech recognition and machine learning, the app coaches the student on pronunciation. Google’s internal analysis shows that after reading 100 minutes on the app, beginner readers—those who read at a speed of less than 45 correct words a minute—see an improvement from 38 percent to 88 percent in reading fluency.
Intelligent farming
Elsewhere, AI is helping India’s agricultural sector. For instance, an AI algorithm can help farmers detect where pests and insects will land in a field by checking the direction of winds, helping them optimise where they plant their crops, says Intel’s Lingam.
According to Microsoft’s Srivathsa, AI is removing guesswork involved in agriculture by enabling data-driven farming.
“With the help of machine learning algorithms and low-cost sensors, our data and AI give farmers a real-time view of soil conditions and moisture, enabling farmers to increase productivity and lower costs,” she says.
Google is working on a range of climate-related efforts such as weather alerts, flood forecasting and air quality in India that can provide timely and important information.
It is also expanding its machine-learning-based flood forecasting it launched in 2018 to help combat the damage by equipping those in harm’s way with accurate and detailed alerts.
Microsoft is working with non-profit SEEDS (Sustainable Environment and Ecological Development Society) to protect vulnerable populations affected by climatic hazards by leveraging AI capabilities.
With a cloud and AI-based model that predicts cyclonic activity, the project is helping vulnerable communities from cyclone-prone areas reach safe grounds on time.
The AI model called Sunny Lives has received technical and financial support from Microsoft’s AI for Humanitarian Action grant and was deployed at scale during Cyclone Yaas in Odisha earlier this year.
Source: moneycontrol.com
