Neeraj Kumar is a Machine Learning Educator at the Alberta Machine Intelligence Institute.
He is currently using machine learning algorithms (deep learning, computer vision and natural language processing) to solve significant problems in digital healthcare.
INDIAai interviewed Neeraj Kumar to get his take on AI.
How did you get into machine learning?
I started machine learning during my postgraduate studies at IIT Guwahati. I was fortunate to work with Prof. Amit Sethi (now at IIT Bombay) for my PhD research which focused on developing machine learning algorithms for computer vision problems. In particular, I have developed learning algorithms for resolution enhancement of natural images that have outperformed benchmark deep learning approaches in terms of image reconstruction performance and computational efficiency. I have also received numerous awards for my PhD work, including the prestigious Microsoft Research India Fellowship and the Erasmus Mundus Heritage Fellowship (which allowed me to spend a wonderful semester in France). After graduating, I moved to the United States (and recently Canada) where I applied machine learning to solve clinically relevant problems at the intersection of healthcare and medicine.
Describe a typical day as a Machine Learning Educator at the Alberta Machine Intelligence Institute (Amii).
As a machine learning educator, I develop and deliver content that enables organizations and individuals to expand their AI understanding and gain the knowledge, skills and abilities they need to be successful in adopting AI in their products and services to be. A typical day at Amii includes:
- meet colleagues.
- Brainstorming business problems that can benefit from AI and ML.
- Identifying the knowledge gaps between academic and industrial AI/ML practices.
- Curate resources to educate employees on adopting AI/ML in their business practices.
What were the initial hurdles and how did you overcome them?
My initial obstacle in applied machine learning research was to identify impactful research problems in medicine and healthcare. Here we could leverage AI/ML algorithms to extract meaningful insights from rich and diverse medical datasets, including genomics, medical images (pathology and radiology), personal health records, and longitudinal data. As a result, I have established interdisciplinary collaborations with medical professionals to identify clinically relevant research problems, procure datasets, and secure funding to achieve my research goals.
In my current role, my main obstacle is to understand the knowledge gap between academia and industry to develop user-friendly resources to bridge this gap. I meet regularly with my Amii team to build and sustain Amii’s external educational initiatives, including providing strategic direction and content development/delivery.
You mentioned that you worked with clinical and medical experts to develop machine learning algorithms for individualized survival prediction based on a patient’s clinical data, medical imaging and genetic data. Can you comment on these algorithms?
Many medical tasks require predicting the timing of a person’s future events – e.g. B. the time to death or disease recurrence of a patient. This “time to event” task is similar to a regression—it describes a patient and predicts a nonnegative real-world value (time to death/relapse)—but crucially, the training data for survival tasks contain censored instances that only provide a lower limit at the event time. This means that most survival prediction models do not estimate a single real value, the “time to event”. Instead, common approaches to survival prediction estimate a different quantity – e.g. a patient’s time-invariant risk score (e.g. Cox proportional hazards model) or their 5-year mortality probability (e.g. Gail model) or perhaps the survival of a population distribution (e.g. Kaplan -Meier curves). In contrast, I focus on survival prediction models that compute individual survival distributions (ISDs)—probabilities of survival at all future time points for a given patient. We built ISD models to enable accurate prediction
- Time to death or hospital discharge in COVID-19 patients,
- Time to onset of breast cancer for women enrolled in the Alberta Tomorrow Project, and
- Time to death in cancer patients.
We are also working to demonstrate how we can translate our findings into clinical settings for hospital resource management in emergencies and to provide women with actionable insights to potentially delay their breast cancer onset.
How do you think India is progressing in AI and ML? From this perspective, what would you describe as our strengths and weaknesses?
Several Indian institutes including the IITs, the IIITs and the IISc have strong AI/ML and Data Science programs that contribute to the upskilling of the workforce and provide excellent opportunities to conduct world-class research. Recently I have also seen strong support from the Indian government to fund educational programs on the one hand and to encourage entrepreneurship on the other. It has helped create several startups across the country solving the most pressing problems in various sectors including e-commerce, banking, healthcare and defense. India’s strength lies in its significant pool of talented young working-age professionals who accept challenges, educate themselves and contribute to economic activity through their exceptional work ethic and skills. The recent rollout of 5G infrastructure will accelerate the adoption of technologies in different sectors in urban and rural areas and create new business opportunities in India. More funding for research, education and small/medium business and more efficient government approval processes will help us (read India) to grow tremendously in the future.
Tell us about the ongoing AI/ML and data science research at the University of Alberta and the Alberta Machine Intelligence Institute.
The University of Alberta (UAlberta) has always been at the forefront of AI and ML. UAlberta’s Computer Science Department and the Alberta Machine Intelligence Institute (Amii) are primarily responsible for cutting-edge research and development in both fundamental and applied aspects of AI/ML. UAlberta’s Informatics department has consistently ranked 3rd in the world for AI/ML for the past 25+ years. Major AI-focused research groups at UAlberta are listed below:
- Board Games Research Group: develops powerful search algorithms and game programs such as Fuego, the first Go program to defeat a top human player in 9×9 Go.
- Games research group: deals with the design, analysis and implementation of artificial intelligence technologies suitable for use in high-performance gaming programs.
- Intelligent Reasoning Critique and Learning (IRCL) group: conducts artificial intelligence research on real-time heuristic search, interactive storytelling, and cognitive modelling. Our most recent applications have been in video games. We have ongoing collaborations with the Department of Psychology, UBC Okanagan, Reykjavik University and Disney Research.
- Working group on medical informatics: is involved in a variety of projects in collaboration with many teams of medical researchers/clinicians to develop systems that effectively learn classifiers that make accurate predictions about future patients. We are now looking at different types of cancer (breast, brain, leukemia), transplants, diabetes, stroke and depression.
Amii is one of three pre-eminent artificial intelligence centers in Canada that strives to bridge the gap between world-leading AI research and its adoption in industry.
What do you think will happen to machine learning in the next ten years?
Forecast to be one of the most disruptive technologies in the world, AI will transform our world and transform how we live, work and do business. It will change our health and economic, legal, cultural and social environment. AI will touch every industry over the next decade and have over $50 trillion in economic impact. The adoption of AI in the industry will lead to the development of smart and connected vehicles, smart, responsive prosthetics, smart homes, better and more accurate diagnostics, and the Internet of Things (IoT). AI will significantly impact healthcare, energy, the environment, the digital economy, manufacturing, transportation, finance and more. As a horizontal enabler, AI will impact every vertical industry over the next decade, and hopefully we will soon set standards for how artificial intelligence can work safely and transparently. There will be processes and laws that ensure the safe and ethical use of AI in the future.
What advice do you have for people who want to work in artificial intelligence research? What should they focus on to move forward?
I think curiosity and a passion for learning are important to start a career in the dynamic field of AI research. The AI community welcomes people from all backgrounds, including engineering, science, psychology, medicine, etc., and there are numerous (online) resources available to fill knowledge gaps for a successful transition to AI. One should strive to build up a solid foundation in probability, statistics, and machine learning. Because AI has the potential to impact multiple industries, it is also important to identify whether an individual prefers to advance AI theory or work on a novel application of AI in their area of interest. Working knowledge of Python programming and familiarity with some machine/learning libraries (scikit-learn, tensorflow or pytorch, jupyter notebook, etc.) will put you on the road to success.
Can you please name some major research publications and books that have influenced you?
I’m a big fan of IEEE Signal Processing Magazine as it publishes tutorial-style articles on signal processing research and applications – useful for beginners to understand the latest research trends.