5G will enable a new era of opportunity, says David Bader
Dr Deepak Garg, Times of India
Recently, David Bader visited India to give a keynote talk at IEEE International Conference on Machine Learning and Data Science at Bennett University, Greater Noida. David A. Bader is Professor and Chair of the School of Computational Science and Engineering, College of Computing, at Georgia Institute of Technology. He is a fellow of the IEEE and AAAS and served on the White House’s National Strategic Computing Initiative (NSCI) panel. He was in conversation with Prof. Deepak Garg, Chair, of Computer Science Engineering at Bennett University.
Some excerpts from the Interview, which will help start-ups to take up some ideas for their new ventures
Question: Big data and data analytics have made a huge impact on businesses in 2017, with trends like artificial intelligence and cloud services being used for their advantage. What are some of these trends that will continue and remain relevant in coming years, and what can be the new trends that take enterprises by storm?
Answer: The enterprise is collecting data at massive rates – and we see the power of information for search – for example, Google – and in advertising such as understand shopper trends from purchases and making better recommendations. The cloud services are now available to collect huge volumes of data, and tools provided to make sense of this information. With Big Data and analytics, more accurate artificial intelligence is possible through Deep Learning technologies. But we are just at the early stages of a data economy. As we collect larger and more types of data and build IT systems that can handle tremendous streams of data, we will move towards predictive analytics. For instance, take cybersecurity. Today, system logs may be collected and after a breach occurs, we apply forensics to understand what flaws were exploited, what information was taken, and who took it. In the future, we will be able to detect attacks in real-time and stop them before breaches occur. In healthcare, as we move towards electronic medical records for patients, data analytics promises to give more accurate diagnoses and move us closer to predictive medicine. With the Internet of Things, and a world of sensors in every facet of life, every aspect of life may be improved through data analytics – from living in more energy-efficient ways, preventative repairs of everything from washing machines to automobiles at the earliest detection and before catastrophic failures, and better understanding of people and populations. It’s an exciting time in data science as we have so much potential to provide strategic predictions and real-time analytics that can improve many facets of our lives.
There are so many articles online that talk about dangers of AI for the society. What do you perceive as a tangible threat and if you will like to give some direction on this issue?
Answer: One of the biggest threats is the over-reliance of AI on the society. For instance, much bias is currently held in AI models for financial and legal decisions. Even though these are often touted as “unbiased” AI techniques, often these proprietary systems contain inherent flaws that make profound mistakes. For instance, there is AI software now that assists judges in the courtroom with the sentencing of offenders based upon predicted recidivism. However, no algorithm is “unbiased” – it inherently captures the bias of the programmer and methods. When using AI and algorithms, we must know better the accuracy and biases with the approach. As another example, financial decisions are often based on past behaviours of communities and associations; which may reinforce poverty within populations and not allow for class mobility. The computer science community must find ways within data science applications to quantify the accuracy and biases in approaches that everyone can easily understand. Until then, it is important that we do not over-rely on these automated “data” approaches.
How can one leverage high-performance computing and big data in the Health Care and Pharma companies
Answer: Today, health care is mostly reactive – which proves to be more difficult to treat patients and leads to more expensive treatments. With high-performance computing and big data, we have the perfect confluence to move to predictive medicine – and we know detecting and treating disease earlier, even preventing disease, is far more effective and less costly. Up to now, healthcare information, such as patient data and medicine efficacy, has been stored on paper records – often handwritten – by doctors, and unusable for large-scale studies. As more medical practitioners move to electronic patient records we are optimistic that causal patterns may be found that improve predictive medicine. Even as more and more people can affordably have their DNA sequenced, we may learn how to group people to decide faster which medicines are more likely effective with less adverse reactions. High-performance computing is also helping with medical image processing and planning radiation therapy for cancer diagnoses.
Bennett University has a special Research Group called MISHA (Machine Intelligence for Smart places and Healthcare Applications). Our curriculum offers Machine Learning and Data Science courses. Based on your expertise, how could one leverage technology to equip students to make them Industry ready?
Answer: Nearly every sector that has data can benefit from Machine Learning and Data Science. Bennett University is quite forward looking to offer this curriculum and having research group with a futuristic vision. I was amazed to see the supercomputing infrastructure of the world standard in a young University like Bennett. Many of these skills will help the budding students to solve some of the big problems of humanity. These techniques are already proving useful for better detection and classifications of cancer, early detection of heart failure onset, and learning to prescribe effective and safe treatment combinations for multimorbidity.
Self-driving cars is the buzz in the automotive world. Could you please elaborate on the advancements in AI in this sector?
Answer: Self-driving cars are a major research area, as highlighted by companies such as Google and Uber that are preparing fleets of cars. And I predict that in a few years, we will see more of these self-driving cars in cities, making deliveries, and operating on transit corridors. The challenge of self-driving cars is making safe decisions in real-time when faced with the unexpected. Deep Learning techniques may improve the accuracy for pattern detection and making decisions when faced with the unknown. Other ethical challenges vex the field, such as making quick decisions during an unavoidable crash and choosing the value between potential obstacles or people.
How will 5G change the market dynamics in Telecom over the next few years?
Answer: 5G will enable a new era of opportunity as we are able to collect and transmit data sensed around the world, opening innovation to our creative minds for new ways to improve our human condition. Take our homes, for example. We are just seeing the entry of automated learning technologies, from thermostats to “surround” Wi-Fi, and energy-efficient dryers. Soon, all our devices will be on 5G, allowing them all to communicate with each other and share information. Not only does this give us a super-fast mobile network, but it enables new applications and our products working cooperatively. These devices may better understand our individual patterns of life, and predict our likes and dislikes, and automate more aspects of our lives.
Could AI also be a game changer in the legal system – We’ve heard of companies in the US talk about “Robot Lawyers”. How can AI help in hastening litigation processes – what are the possible outcomes with AI applications?
Answer: While some think that AI could provide “robot lawyers” and automate the legal system, I’m somewhat of a sceptic to trust my legal protections solely to AI! Certainly, data science can help scour the case histories for related information and legal claims, and reduce fraud within the legal system. However, AI relies on models – and it will be some time before we know how to create truly “unbiased” decision systems, and if even at all possible! So, until then, I hope we are cautious with placing our fate in the hands of robots.