The human intelligence behind intelligent machines

Dr. Sanghamitra Bandyopadhyay is a professor of machine intelligence and director at the Indian Statistical Institute. She is the first woman to hold this position. She has also been a member of the Science, Technology and Innovation Advisory Council of the Prime Minister of India—one of the country’s highest honours for a scientist. 

The Infosys Prize in 2017 in Engineering and Computer Science for her work on algorithmic optimisation in biological data analysis, and the Shanti Swarup Bhatnagar Prize in 2010 for outstanding research in Engineering Sciences, are among the multiple awards in her illustrious career.

Lauded as an inspiring example of original research in computer science done entirely in India that has had worldwide impact, her discoveries include a genetic marker for breast cancer, determination of co-occurrence of HIV and cancers and the role of white matter in Alzheimer’s disease. 

But of her numerous achievements, Dr. Bandyopadhyay has said, “What makes me happiest is that my son is so proud of my achievement.”

I learnt so much from her about:

1. The big and complex problems that she believes machine learning can help solve

2. Why silos hinder breakthroughs in science, and what we can do to diffuse them

3. How to develop an innovation mindset and pave the path for more women into STEM

What gives Prof. Sanghamitra Bandopadhyay her “hustle fuel”?

Highlights from the conversation

Data without a sound approach can generate more noise than signal. 

“Often we think that if we train it (the algorithm) with tons and tons of data, everything will be fantastic. That need not necessarily be true… It’s not just about how many data points you have, but also how many features.”

Dr. Bandyopadhyay eloquently described machine learning as the subfield of artificial intelligence that gives computers the ability to learn without explicitly being programmed. Traditional programming similarly requires creating detailed instructions for the computer to follow. Machine learning takes the approach of letting computers learn to program themselves through experience. 

A common assumption people have is that more the data, the better trained the program. Probably one of the most famous quotes defending the power of data is that of Google’s Research Director Peter Norvig claiming that “We don’t have better algorithms. We just have more data.” However, Dr. Bandyopadhyay emphasized how significant value lies in tweaking the data one trains their algorithms on, including sampling and extracting core features in the data, to help push it toward more accurate and predictive results.

Artificial learning and machine intelligence hold tremendous potential to address pressing problems, but we need to prepare for a continually evolving paradigm.

“That’s what human beings are experts in—they can plan, they can strategize. That’s also where artificial intelligence/machine learning will play a role—they will help the human beings in making these decisions.”

Given the amount of data being generated and the need for machines to be intelligent and generalize to new situations, Dr. Bandyopadhyay suggested that applications of AI/ML in domains ranging from healthcare to climate change, e-commerce, and the design of effective government policies, is here to stay, but there is still work ahead to harness machine learning’s full potential.

When considering the ability of quantum computers to aid in machine learning in the coming years, Dr. Bandyopadhyay highlighted the importance of developing methods that will work on systems beyond existing binary systems. 

Machine learning has the biggest impact when it’s developed by cross-functional teams with a mix of skills and perspectives. Moving from siloed work to interdisciplinary collaboration requires deep mutual trust and respect.

“People from different areas of expertise speak very different languages… A biologist needs to understand a little computer science and mathematics, and a mathematician/computer scientist/physician/statistician needs to understand a little biology. But that does not make them a biologist. That is where mutual respect is of utmost importance.”

When solving complex interdisciplinary problems, a working understanding of the different relevant fields is important. Through her explanation of how she worked to identify a genetic marker for breast cancer, Dr. Bandyopadhyay illustrated how she herself is a computer scientist, but understands biology well enough to be able to work with multidisciplinary teams to drive to the answers and have impact. But she emphasized how success ultimately rests on people with different expertise working together with mutual trust and respect. 

Nurture the innovator’s mindset from a very young age through example, iteration and reflection.

“At a very early age, there should be fun learning. Up to Class 2, it’s not necessary to teach at all—it should just be learning from nature, learning to observe nature… Get that observation into the student, they don’t need to learn anything from books.”

Dr. Bandyopadhyay highlighted how important it is for children at a very young age to explore, tinker, discover and create. Her belief is that unstructured and creative exploration at an early age can make the serious seem fun, and make the young innovator more enthusiastic about STEM, well before they are confronted with the  pressures and challenges that invariably lie on the road ahead.

Aside from hands-on exploration and learning, she was also a big advocate for young people developing a hobby outside their academic and professional pursuits (as she did when she was young, with her interest in music, cinema and reading). Listening to her talk truly highlighted how being well-rounded is key to having a mindset of inquiry, imagination, problem solving and passion-based learning.

(Hustle Fuel represents my own personal views. I am speaking for myself and not on behalf of my employer, Microsoft Corporation.)

About the Hustle Fuel series: Hustle Fuel, a Founding Fuel series, looks at the world of work and entrepreneurship from a woman’s lens. Building a company or a meaningful career is brutal, and role models for a path less trodden are always invaluable. The Hustle Fuel series is relevant especially for women—but not just for women. 

Thriving in the evolving workforce demands ‘hustle fuel’. It demands having to punch above one’s weight to earn a seat at the table—not because you are a woman but because you are the right person for the job. Interestingly, it just so happens that this hustle fuel is precisely the attitude any entrepreneur needs to survive. Whether a man, woman, or from an ethnic minority community.