Babak Hodjat: Inventor of Natural Language Technology Used in Siri
With the goal of harnessing the untapped potential of Iranian-Americans, and to build the capacity of the Iranian diaspora in effecting positive change in the U.S. and around the world, the Iranian Americans’ Contributions Project (IACP) has launched a series of interviews that explore the personal and professional backgrounds of prominent Iranian-Americans who have made seminal contributions to their fields of endeavour. We examine lives and journeys that have led to significant achievements in the worlds of science, technology, finance, medicine, law, the arts and numerous other endeavors. Our latest interviewee is Babak Hodjat.
Dr. Babak Hodjat is co-founder, chief executive officer, and chief scientist of Sentient Technologies, creator of the world’s most powerful distributed artificial intelligence platform. A serial entrepreneur, he has started a number of Silicon Valley companies as main inventor and technologist, and he is well-known as the architect of the technology behind Apple’s Natural Language conversational system, Siri.
Babak is the former senior director of engineering at Sybase iAnywhere, where he led mobile solutions engineering, and before that, he was co-founder, chief technology officer, and board member of Dejima Inc., where he was the primary inventor of its patented, agent-oriented technology applied to intelligent interfaces for mobile and enterprise computing — the technology behind Siri.
Babak is a published scholar in the fields of artificial life, agent-oriented software engineering, and distributed artificial intelligence, and has 25 granted and many more pending patents to his name. With an unmatched level of expertise and enthusiasm, he equips audiences with insights on the future of artificial intelligence, including the impact of natural language processing, machine learning, genetic algorithms, and distributed AI on how business will be done.
Tell our readers where you grew up and walk us through your background. How did your family and surroundings influence you in your formative years?
I was born in London, went to kindergarten in the US, went to middle school in London, finished high-school and undergrad in Iran, and did my PhD in Kyushu University in Japan. My father is a retired university professor in Entomology and my mother paid a lot of attention to her children’s upbringing and education, so, of course, I owe a lot to them. I think, from my dad, I learned to view the world from the lens of scientific inquiry, and with his words of encouragement, I got a heightened sense of self belief. From my mom, I learned critical thinking, perseverance, and the love of books.
While in London, in 1978, my father worked at the British Museum. As an eleven-year-old, I didn’t particularly share his enthusiasm for creepy crawly insects, and so, when visiting him, I would wander out to the nearby Science Museum, where they had an actual computer on display on the sixth floor. I would spend long hours there, trying to figure out how it worked, and that memory stayed with me years later when deciding what to study in college.
Later on, the experience of growing up in the aftermath of a religious revolution encouraged me to think about the nature of things: Where did we come from? What is the nature of intelligence? What is life? Can they be recreated in computers? It is a quest I am still on.
What is it you really hope to accomplish with Sentient Technologies? Either directly with your company or indirectly influencing the whole field in AI over the next decade?
AI has a lot to offer to the world, much beyond its current use cases. I would like to prove the viability of AI as a widely applicable technology. We are at the dawn of a revolution on par with Computing and Software itself, and Sentient is well positioned to take part in it.
Over the next decade, I would like to help introduce AI technologies well beyond the current state of the art, in many different fields.
You believe humans are too emotional for the stock market. Therefore, you started utilizing Sentient’s artificial intelligence platform to develop, evolve and optimize its investment strategies in finance. Could you elaborate on this for our readers?
Humans are appropriately emotional, but for the types of environments we evolved to survive, which do not necessarily correspond to trading in the stock market. An AI-based system can be trained and assessed on much more data, and at timescales that are impossible for humans to operate. Furthermore, all of an AI (or any automated) system’s actions and context can be logged, reviewed, and interrogated. Something we cannot easily do with human traders. Finally, human decision making and thinking is often subject to our context, and our emotional state. As humans, we abstract our context, along with our physical reaction to our context, and give it a label that is often categorized as an emotion. This abstraction then influences how we react to the world, in ways that we might not be conscious of. This bias generally serves us well in dealing with our environment and society, but might not be easily rationalized for problem solving in less familiar environments that require more consistent behavior.
Can you tell us the significant trends in artificial intelligence and how companies can deal with them?
The recent resurgence of interest in AI has been fueled by breakthroughs in Neural Networks called Deep Learning. These technologies are particularly good at modeling systems for which we have lots and lots of labeled data. They have primarily been used to sense the world (e.g., speech recognition, vision), but, often in conjunction with other AI techniques, Deep Learning systems have shown promise in other application areas such as Machine Translation, or playing games like Go, Atari, and Chess. Of course, there have been other breakthroughs in scaling AI systems beyond toy-problems, including in the field of Evolutionary Computation, which my company, Sentient, specializes in.
In the past decade, cheap computing and storage, and increased digitization, have meant that companies have been able to collect and store lots of data relevant to their business. The trend now is to see if, by applying AI, they can find models and insights from this data that can help their respective businesses.
I would caution, however, that building AI systems is hard. There is no one-size-fits-all solution, and much knowledge engineering and AI expertise goes into tailoring an AI solution for a given domain. Just like the early days of the computing and digitization revolution, companies should be prepared to invest in AI talent and consult with experts on their AI needs and road-map.
Another field of AI which is enjoying a lot of attention is conversational systems. This started with the success of Siri, but many companies are interested in chat-bots as a new way to interact with their customers.
My word of caution here is that, ever since Eliza, it has been easy to fool people to think that a conversational system is much smarter than it really is, and considering these systems as a substitute for humans has proven a challenge.
What is the biggest challenge that you face in your career?
So much to do, so little time!
Could you share with us some of your ongoing work in your areas of interest, present some of the opportunities and challenges in these areas, and forecast some future directions and possibilities?
One fascinating and widely applicable area we are focused on is Black-box Optimization through surrogate modeling, where a model of the real world application domain is built, typically using Deep Learning, and then repeatedly interrogated, using evolutionary computation, in order to construct a decision making solution. The solutions are then tried out in the real world, and the outcomes help improve the original model. This is especially effective for systems where experimenting in the real world is expensive or time consuming, and so, effectively, the AI is driving the data collection process in an intelligent way. This approach is applicable to fields as diverse as Cyber-Agriculture, and optimizing software development, testing and maintenance processes.
Another area of interest has to do with automatically engineering and designing Deep Learning solutions, using a different AI technique called Evolutionary Computation. Building elaborate deep learning systems for different domains is hard work and often takes many PhD-months. Evolutionary AutoML, as this technique is called, is analogous to having AI design the AI.
Ultimately, I would like to build a system in which a population of actors, sometimes competing, sometimes collaborating, solve problems in a robust and decentralized manner, as an emergent byproduct of surviving a problem’s environment. We are still far away from this goal, but I believe it is doable, and the next essential step in the evolution of AI systems. This is a sub-field of Artificial Life called Applied ALife.
How far has artificial intelligence advanced over recent years and what impact has it had on healthcare so far?
I am not the best person to answer this question, as I’ve not been focused on applications of AI in health-care recently. My limited understanding of the state of AI in health-care is that, so far, AI has primarily had an augmentative role, helping practitioners with their decision making. For example, advanced AI-based systems can now identify diseases from medical images with high accuracy. There is much promise in this field, however, ranging from the role of advanced robotics in surgery, to predicting the onset of septic shock in ICU patients, to improved health monitoring via devices like FitBit or AppleWatch.
How do you see your field changing? What excites you most about the future of AI?
The wide scope of interest in AI-based solutions is a dream come true for me. I have been working in this field since the late eighties, and it is finally being taken seriously outside of scientific and research circles. The more people work in the field and build AI-based practical applications, the more the field will progress. The world is faced with daunting and complex problems, and our most advanced tool with which to tackle these problems is AI.
In your view, what is the biggest challenge with which your field is currently grappling?
Through its history, AI has over-promised and under-delivered. This is because there is a deep chasm between popular notions of AI, rooted in science fiction, and the reality of the state of AI. I think we need to educate people so that society’s reaction to AI is proportionate to the reality of where it is and its promise, rather than disproportionate to the perceived threat of Science Fiction AI.
The other threat is in focusing too much on one AI technique (i.e., Deep Learning) at the cost of neglecting all the other fields of AI.