Big problem? Keep calm and let Google AI provide a solution for it
Artificial intelligence or AI has become a buzzword these days. This technology today is being used everywhere from social media sites for detecting and removing illegal and inappropriate content to cameras in our smartphones for processing images. While all these are common use-cases where AI is being deployed aggressively to solve the issues at hand, there are other areas like healthcare and environment conservation where companies like Google are making good use of their expertise in AI and machine learning.
Google on Thursday, July 10, 2019, held Solve with AI conference in Tokyo, Japan with an aim to highlight how AI solutions can provide solutions for problems that are too big to be tackled easily. The company showed that how it is working to help researchers solve some of the most difficult social and environmental problems, including in areas such as healthcare, environmental conservation, disaster prediction and agriculture. The day-long event started with an opening keynote by Head of Google AI Jeff Dean.
"We believe AI can help tackle some of the most difficult social and environmental challenges of our times, and not just in computer science but in areas where you necessarily expect it like healthcare, environmental conservation and agriculture," he said as he explained the basics of AI and the efforts made by Google towards solving some of these issues.
Dean's opening note was followed by a series of presentations by Google product managers in which they highlighted the way Google was using various machine learning (ML) models to solve big problems. First in the list was healthcare.
Early detection of a disease is one of the key factors that ensures that the patients have a better prognosis. And so, early detection of diseases, such as cancers, would ensure that patients have a better quality of life. At present, Google is using its ML models for helping doctors detect three diseases - lung cancer, breast cancer and diabetic retinopathy - at an early stage and treat the patients better.
While in case of lung cancer, Google's AI researchers train the ML algorithm to analyse CT scans and predict lung malignancies, in case of breast cancer, the ML algorithm is trained to detect lesions in the samples collected by the doctors. And in case of the diabetic retinopathy, AI researchers trained their ML systems to read retina scans to detect the condition before it gets to a stage where a patient has to undergo a surgery. The company has also partnered with several hospitals in India to help doctors detect this condition.
Apart from healthcare, Google highlights how the company was using AI to predict floods using a variety of factors including satellite images and the data pertaining to the flow of water in a region. The company piloted the program in Patna last year and it is now planning to expand the scope of this tool to monitor to the rest of the country and then to the world.
Decades of hunting the wild animals have put many of these species - such as the humpback whale - at a risk. While conservationists try to track these species by listening to their songs using under-water microphones, such as National Oceanic and Atmospheric Administration (NOAA) has collected over 19 years worth of data, the amount of data collected makes it impossible for researchers to track them accurately. Now Google is using its ML algorithm to identify these whale species in the long underwater recordings.
As mentioned before, Google is also using AI to help people with hearing disability or people with age-based hearing loss to communicate with their friends and family better. The tech giant has developed an app called Live Transcribe that uses speech recognition to automatically transcribe conversations in real time, allowing people with hearing disability to read the contents of an ongoing conversation on their smartphones.
Google's Live Transcribe app can detect 70 languages and it is available in 150 countries across the world.
Impressive? Yes, it is. But there are more use cases.
Google's machine learning algorithms are also being used for other purposes such as for transcribing ancient Japanese scripts, tracking illegal logging in rainforests and for detecting pests on crops. While some of these issues have been handled by Google directly, others are being taken up by organisations like the Rainforest Connection, Gringgo Indonesia Foundation and the Wadhwani AI Institute and are being funded by the tech giant as a part of its Global Impact Challenge 2019, in which 30 organisations share a grant of $25 million from Google.org. In addition to money, these organisations also get mentoring from Google's AI experts to improve and scale their products to tackle the big challenges.
The writer's travel and stay in Tokyo, Japan have been arranged by Google.