As a young girl growing up in India, Huma Abidi was hooked to science fiction and fantasy. Back then, she thought she would be a doctor, but computer science was where her destiny lay. Today, she’s the Director - Machine Learning/Deep Learning Software Engineering at Intel Corporation and is responsible for deep learning framework software optimization for Intel Xion processors.
She was named by 'AIthority' magazine amongst the 15 women using AI to change the world. At the recently held Artificial Intelligence 2018 Conference in Beijing, she led a session on Optimizing Deep Learning Frameworks for Modern Intel CPUs.
Read other Stories in #WomenoftheDiaspora
With a BS in pre-med and chemistry, followed by an MS in computer science from the University of Massachusetts, she joined Intel as a software engineer and has since risen through the ranks. Apart from AI, she is also passionate about empowering underprivileged girls and is on the board of ROSHNI, an organisation that educates and trains underprivileged girls in India to become financially independent.
One advice that I got from a senior woman executive was that if you could not find women managers, find a man who has daughters — he would want you to move forward in your career, because he wants to see a world where women succeed and I did.
SheThePeople.Tv spoke with Huma about her journey into AI, being a woman in a very male-driven industry, the challenges AI presents as well as the opportunities to humans, and how we can get more young girls to take up technology as a field of study.
As a little girl, what did AI mean to you? How did you get attracted to this field, and were you ever fighting pre-conceived notions about Tech not being a space for women?
When I was growing up in India, AI wasn’t really on my radar. I was always interested in science, technology and science fiction, which was more of a fantasy in the form of comic books or the Star Trek series on TV. There were much fewer women in engineering back then, and if you were good at science, medicine was a clear career path for women. My goal was to become a doctor, so I enrolled in a Pre-med and Chemistry major. I credit my mom who enrolled me into an extensive computer programming class for really getting me interested in technology.
Tell us a bit about your journey as a woman in tech and AI. What were the biases and apprehensions you had to face in the professional space, implicit or overt, through the years?
In the early days, like 20 years ago, opportunities for women in technology, especially at senior level were very few. It was a different atmosphere: more often than not you’d be the only woman in meetings and it seemed easier to remain quiet than to make sure your point of view was always heard. Almost every year at my performance review I would be told I was “too nice”, that I needed to be more aggressive or forceful to be seen as a leader. And my response would always be the same: that I didn’t want to change who I was. I was fortunate enough to find some great managers and roles models, both men and women, who helped me throughout my career. One advice that I got from a senior woman executive was that if you could not find women managers, find a man who has daughters — he would want you to move forward in your career, because he wants to see a world where women succeed and I did. I have also mentored and helped several women who were facing issues at work, e.g. not getting promoted, by coaching them and often connecting them to the right folks.
Almost every year at my performance review I would be told I was “too nice”, that I needed to be more aggressive or forceful to be seen as a leader. And my response would always be the same: that I didn’t want to change who I was.
Things are very different now. There is a big push towards equality and all tech companies are working hard to close the gender and diversity gap. Surveys show that 15-28% of women hold tech jobs at tech companies polled. It was a proud moment for me as a woman, as an Intel employee and an attendee at the Consumer Electronics Show in Las Vegas in Jan 2015 when Intel CEO announced a new diversity in technology initiative of investing $300 million with the aim of achieving “full representation” of women and under-represented minorities at Intel by 2020.
Since then I have seen numerous efforts throughout Silicon Valley to hire and to retain women at different roles, by providing them guidance, mentorship and opportunities to succeed.
Have things changed since, what is the representation of women in a space as narrow as AI? When I read a list that says Top Women in AI, I wonder if we should be encouraging these lists at all, or working towards a situation where the gender ratio is so equal that they make such lists redundant. Your thoughts?
AI is a relatively new area with recent breakthroughs, I would say the challenges and progress regarding representation of women is probably the same as technology in general.
Regarding the list, the two don’t have to be mutually exclusive — while we are working towards a situation where the gender ratio is equal, it’s important to highlight female role models along the way. If young girls only see men in charge, they may feel that they don’t fit the mould of an engineer, and they may feel out of place or intimidated. If there are senior women in AI, it might be easier for young women to see themselves in such a role.
If young girls only see men in charge, they may feel that they don’t fit the mould of an engineer, and they may feel out of place or intimidated.
A couple of weeks ago, I was a judge at an AI / Deep Learning Hackathon event in Hollywood LA. Though the ratio was quite low in comparison, I was quite happy to see a few female hackers. One young lady made it a point to come and talk to me twice, and tell me how great it was to see two women judges at such events, and that we were role models for students like her.
How do you think we can get more girls into STEM streams, how do we counter the pernicious narrative about girls not being good at Math and Science that seems to deter them?
This is a problem all tech companies are facing: unless we build a pipeline of women in technology, the gender gap will not be narrowed or closed. The change needs to happen early on before young girls start developing negative perceptions: they need to know that it is cool to be a scientist or an engineer. Science and Math needs to be made interesting both at school and at home. Tech companies are providing Science grants to schools to add technology classes, which are meant to show young children that science goes beyond a subject, and how it is part of daily life. There are now many orgs such as Girls Who Code, Girls Innovate, Girls Geek and many more that get young women involved in Computer Science and open up a wide variety of possible career paths. There are very few TV shows or toys that are designed to promote science especially to young girls. I have spoken to an audience of girls ranging from second grade to high school to talk about why I chose the career I did, and these are talks arranged by parents. My daughter and my son both chose engineering and Computer Science as their majors and I think we may have nudged our daughter towards CS more than our son.
My daughter and my son both chose engineering and Computer Science as their majors and I think we may have nudged our daughter towards CS more than our son.
Tell us about your work at ROSHNI. What was it about this NGO that made you want to associate yourself with it?
As a woman who grew up in India, ROSHNI’s mission is very personal for me. Giving a young woman an education has the power to change lives, and even empower entire communities. I got involved with it because of my connection to its founder Saima Hasan, who after graduating from Stanford, moved to India to establish Roshni in 2008. By providing employability and life skills training, ROSHNI aims to break the cycle of poverty. Over 80 per cent of Roshni alumni are pursuing higher education along with part-time employment, and many have been awarded academic scholarships through Roshni. I have been involved with Roshni since 2012 and serve on the board since 2015.
I’ve been reading articles that state China will win the race for AI by 2030. Would you agree? Where do you think AI is headed in countries like India?
It is true that the Chinese government is investing heavily in AI, But there is a lot of exciting research going on with Machine Learning in all areas of life and in all countries. Hopefully, this is not a race and everybody will benefit from it. In April, I was invited to the prestigious O’Reilly Conference on Artificial intelligence in Beijing to talk about my team’s effort in enabling AI software to run fast on Intel Hardware. I found that people, and media in particular, were very aware of the benefits as well as the pitfalls related to AI. Due to a lack of enough trained healthcare professionals in China, there were several discussions on how Deep Learning can be used to detect diseases. On the other hand, they expressed concern about potential misuse, privacy invasion, fake news, etc.
India is well known as a tech hub, and I see it taking full advantage of advancements in AI as it relates to health, farming, transportation logistics, fraud detection, autonomous driving, etc. However, there is great potential for misuse of AI, such as creation of fake audio/video (the famous fake Obama footage) which is practically indistinguishable from real. This is very dangerous to a society which is often seen as somewhat gullible/ not very tech savvy as a whole.
India is well known as a tech hub, and I see it taking full advantage of advancements in AI as it relates to health, farming, transportation logistics, fraud detection, autonomous driving, etc.
With innovators like Elon Musk sounding the war cry against AI, do you think we need to be worried about AI taking over our lives in a dangerous way rather than the good way we envisaged?
As mentioned above, there is potential downside to all technology, but that shouldn’t stand in the way of progress and learning. Regulations will play an important role on how to control the use of technology. I remain optimistic, because right now I see smart machines augmenting human intelligence. Intelligence has many forms, and even with a high IQ, machines will not have the emotional intelligence that human being do.
Would you agree that AI is the new atheism? Do you believe there is a case for a Human-In-The-Loop in AI?
I believe in having checks and balance. In general, there must be human control over decisions, guided by ethics. If and when the technology is fool-proof, and has been well-tested, then it can be deployed autonomously in some domains. And there are domains where AI should NOT be deployed autonomously, one such example is of Chicago Police using AI to predict crime patterns which resulted in them heavily racially profiling certain areas. US law enforcement has due Process and requires reasonable cause for invasive policing. AI should not be considered reasonable cause. What can we do about it? These are the questions that need to be answered. The assumption that AI is unbiased is simply not true because the data used may have stereotypes and the program will be trained with the same biases.
The assumption that AI is unbiased is simply not true because the data used may have stereotypes and the program will be trained with the same biases.
How can we, as regular everyday people, prepare ourselves for the quantum revolution that AI will bring to our everyday lives?
I don’t think “the revolution” will happen overnight – the infrastructure replacement will take time, even if technology is there, which we could use to futureproof ourselves. First of all, AI will optimize and further automate manual jobs. As in the industrial revolution, the change will be in the kind of jobs that humans will do. We will need to retrain ourselves to the jobs of the future: education, philosophy, and more importantly government policies will need to change.
Workers in all sorts of jobs will have to consider how to leverage the new technologies and tools for their benefit.
For example, AI algorithms can bring automation to even a highly-skilled, highly-specialized field like Radiology. This can offload routine, time-consuming tasks from human Radiologists, lead to improved diagnoses, and ultimately benefit both the medical professionals, and their patients.
Similarly, there are use cases in so many different domains: In farming, for detecting crop diseases; in Banking, for fraud detection; In factories, and in mining, for accident prevention. In all these cases, AI offers powerful and useful tools to augment the labor, intelligence and creativity of humans.
And finally, in India, AI is not perceived as a very hot field, there doesn’t seem to be much interest in it, according to a recent survey of most sought after professions. What would you attribute this to, and how do you think this could be changed?
Honestly, I am surprised to hear that, but I think this will change quickly, because now we have use cases for AI and how it impacts the world. It has moved from theory/technology to use case. The pace of innovation in this field is very rapid and very quickly India will catch up in investing in AI. The fact that the cost to learn Machine Learning is literally free because of Coursera and other MOOCs (Massive Open Online Courses), there is no reason that India would not gain expertise in AI very quickly.
Kiran Manral is Ideas Editor at SheThePeople.TV