Machine Learning in Practice

Interview with Behnaz Abdollahi

Behnaz Abdollahi is a machine learning engineer at biotech startup Gen-9 Inc, working on developing machine learning algorithms on wearable devices. One of her project during her PhD studies was applying machine learning algorithm to automatically rank vascular tissue performance with focus on breast tumors. The principle strength of the method is ranking tissue vascularity independent of the tissue type. Behnaz focuses on developing algorithms in different domains such as wearable devices, healthcare and image analysis.

We look forward hearing more about Machine Learning in Practice from Behnaz at AI With the Best online conference April 29–30th, and are pleased to have had a chance to interview the Machine Learning Engineer beforehand.

Q What personally motivated you to begin your work in machine intelligence?

I received my bachelor’s degree in computer engineering, which helped me to acquire basic knowledge of computer engineering, machine language, programming and applied mathematics, and highly motivated me to pursue advanced research in the general area of computer science. The first piece of code that I wrote as an undergraduate student made me feel amazing. I remember feeling that I had created something that can work automatically and independently by processing information and responds to unseen input logically. That is intelligence at a very basic level, but there is so much more to it. Intelligence also involves learning from past experiences, collaboration, parallelization, self-awareness, prediction of actions and states, self-sustainability and many more characteristics. I wanted to know what the limit of machine intelligence is, and how it can be pushed to achieve unimaginable tools that improve human beings’ quality of life.

I chose machine learning as the topic of my masters and PhD research and focused on biomedical and healthcare application. During the PhD studies and after graduation, I have developed several computer vision and machine learning algorithms to improve the analysis of healthcare datasets.

They are all based on machine intelligence: pieces of software capable of autonomously and efficiently processing information and making optimal decision. It gives me the utmost enthusiastic to realize that the smart technologies that I contribute directly improve the quality of life and well-beings of people.

Q What do you personally find most exciting about your current role?

My current role involves developing machine learning algorithms for wearable devices to help elderly people. I build machine learning models that can be embedded on smart devices to help make the day to day life of a group of people safer and more convenient, and that is extremely valuable to me. On top of that, the fact that I can directly apply my expertise, knowledge and research experiences in practice and get to observe the progress and outcome is intellectually very rewarding.

Q How has the field of machine learning changed since you began working in it?

I started my master’s studies in artificial intelligence in 2003 and that was probably the first time I started to explore this area. Since then, the technology of computation in general has grown significantly and is now much more advanced. Specifically, advances in computer hardware, parallel programming, cloud computing, and distributed computing technologies have made it possible to execute very intensive calculations that often rely on extremely large amounts of data. Therefore, many machine learning tools that were previously only considered “theoretical” or “asymptotically optimal” are now practical tools that are being revisited, studied more concretely and designed more efficiently and actually used in practice. Deep belief networks are one example of such tools that is the building block of deep learning. In summary, machine learning is now more than ever entangled with the science of big data and distributed computing.

Another big change that I noticed is the industrialization of artificial intelligence. Big companies like Google, Microsoft, Amazon, Facebook, etc. have invested significantly in machine learning in the past two decades, and thus, this field is now a lot more known to the public and is well beyond only academic research. It is a lot more tangible in our daily life too with various machine learning based apps on smartphones, powerful search engines, recommendation systems, social networks, self-driven cars which are yet to be realized and so on. To summarize, we are dealing with machine intelligence in practice a lot more often and a lot more closely than say 10 years ago.

Q How do you ensure collaboration and smooth organisation in your team?

Every member of a team has a unique talent, unique problem solving approaches and a preferred way of communication. I make sure that I understand and appreciate those differences and make the best use of it. It is very important to create a professional environment in which every person has the freedom and is in fact encouraged to think and work on the pieces of a project independently and innovatively, and at the same time be able to share thoughts and respect the team work spirit when needed. In my opinion, the most important factor to guarantee team organization is to make sure that personal incentives are aligned with the benefits of the group as a whole. It is very important to want the best for everyone, and to make sure that the people you work with have good intentions in addition to being capable and dedicated.

Q What advice would you give to budding AI developers?

Most important of all, enjoy what you are doing. This is your life and don’t be afraid to embrace it and learn it. Creation comes with passion, working hard, and pushing the boundaries without fear. I believe that machine learning is not my career, but my passion.

Q Are you excited about speaking at AI With The Best?

This year is the second year that I am speaking in the “AI With the Best”. Last year was the first time I spoke to a diverse group of people with different backgrounds publically, and it was a great experience. The organizer told me that my talk was one of the popular ones in the last year, so I am really motivated to prepare a useful talk for this year as well.

Thank you!

You can ask Behnaz your own questions, and learn more in her talk about using Machine Learning in Practice at our upcoming AI With The Best, Online Developer Conference 29–30th April.