No prizes for guessing the trending topic in technology today. Every company — from Microsoft and Amazon to your next-door few-month-old startup — is trying to make the best of Artificial Intelligence, or AI.
AI is finding application in many areas. For example, healthcare (prediction of occurrence of diseases going by past health records, data from wearables, demographic details); logistics and supply chain (to estimate package delivery date and time going by orders, current location of package, sensor data); identification of customer movement patterns inside a department store (going by video recordings from within the store, transaction history) and how self-driven cars should park and avoid traffic (going by data from sensors of self-driven and other cars).
“Decision-making can be abstracted into three steps: pattern recognition, interpretation of pattern and taking suitable action,” says Shashank Dubey, Co-founder and Head of Analytics, Tredence, an analytics services and solutions company serving some of the leading Fortune 500 clients. “All three can be automated to some extent using machines. The advantage of using machines is justified by the time saved for the experts — who will be involved in supervision only, instead of the whole decision-making process,” he says.
Keeping pace with technology
Deployment of new-age technologies, like AI and Internet of Things (IoT), brings with it enormous challenges, requiring service providers to become more adaptable, says Dubey. Between 2013 and 2017, the team size of Tredence grew from six to over 250, and the number of clients went up from two to over 24. It was recently recognized by Inc. 500 as one of America’s fastest growing private companies.
New-age technologies have also changed the way companies interact with clients. The era when they came up with a product that resolved the problems of a client is fast giving way to one where both sit together and jointly evolve a solution. Says Tejinderpal Miglani, CEO, Incedo, a technology-solutions provider, “Traditionally, vendors approached clients from a billing or a revenue point of view. The onus of making a solution work was left in the hands of the client. However, the trend now is to approach from the other side.”
Incedo recently launched in Bengaluru its IoT NXT Lab, which provides a simulated environment for collaboratively working with clients to ideate, design and develop end-to-end IoT solutions. Real-time data from interconnected sensors and devices are put through cognitive platforms and predictive analytics, to get actionable insights.
Treatment of cancer
One of the fields where AI is making an impact is precision medicine in the healthcare industry. It refers to tailoring of medical treatment to a patient. It’s finding increased application in oncology since every tumour is different and hence the same treatment does not work for all patients. “There is technology available today to measure thousands of tumour characteristics (big data), but there is a lack of know-how to use that big data for making clinically informed decisions,” says Taher Abbasi, Co-Founder & COO, Cellworks Group Inc., which uses precision medicine technology to address the needs of patients, physicians and pharmaceutical companies.
The technology involves identifying optimal therapies for a patient, or novel therapy options where no standard option currently exists. A digital computer model of a patient is created which integrates “the disease-biology knowledge” with the patient’s unique characteristics. The right therapeutic response is determined by using simulated digital models of therapeutic agents, and mapping the right therapy to the right patient.
Explaining the progress made, Sri Krishnan, Vice President, Robert Bosch Engineering and Business Solutions, says the application of AI in driving has been evolving over the years. “The first stage is in assisted driving and parking. It involves use of cameras to ensure that the car stays within the lane; any deviation from the path is indicated by a warning beep,” he says. The technology, which has been in operation since 2010, is intelligent too: if the car is overtaking another vehicle for example, the sensor doesn’t beep. The second stage in the evolution is ‘partial driving’, which involves switching between automated and manual modes. “When it is on highway, the car tells the driver that he can switch to automated mode. Taking cues from cameras and sensors, the car alerts the driver when there is a need for him to take control of the vehicle. Krishnan says we are only in the first two stages, with the next two — highly automated and fully automated — still a few years away.
Businesses communicate with lakhs of customers. It is imperative that the messages (emails or notifications), are relevant to the customer. Bengaluru-based startup Boxx.ai says it has democratised analytics so that customer engagement becomes more relevant through AI. Prakhar Raj, Co-Founder, Boxx.ai, says, “The product does two things. One, it uses historical data to predict the probability of any customer buying a particular product; two, it is able to communicate the most personalised and relevant products to each individual through multiple channels, like emails, app and browser notifications, and Facebook and Instagram ads.”
Transformation of the GPU
Developments in Artificial Intelligence has led to the transformation of GPU (graphics processing unit), which was originally invented for immersive 3D graphics in gaming.
“AI will not be an industry — it will be part of every industry. We already see a strong focus on security, intelligence and investigative capabilities. This includes advanced search and facial recognition analytics, made possible by AI-powered computer vision,” says Vishal Dhupar, Managing Director, NVIDIA — South Asia.
Deep Learning involves a number of complicated mathematical computations. GPUs have become so crucial today because they have processors comprising thousands of cores which can do millions of mathematical operations simultaneously.
Is AI a threat?
There are fears that machines, trained to mimic humans, might just run amok and take automated decisions which might be dangerous for all of us. However, most industry experts say we are at a very early stage, and any such dangerous possibilities lie far ahead. Dhupar says, “There are complex questions surrounding this field, from data privacy and transparency into how artificial intelligences are solving problems, through to accountability and liability. Technology providers will work together with industries and governments to address these complex problems and ensure the value of this transformative tool is realised responsibly.”