A member of the 2014 class of IBM Fellows, Dharmendra is excited about inventing a new computing architecture. “This honor cements the mandate to make [SyNAPSE] successful!” To which he then stressed that the success of the SyNAPSE Ecosystem, and his being named to IBM’s top technical rank, has everything to do with his brilliant colleagues at the Almaden research lab in San Jose, CA.
Cognitive computing started as a pet project for Dharmendra in 2005. It led him to chair the eponymous Almaden Institute in 2006. But taking on a complete reinvention of modern computing (not to mention the 100 trillion synapses in the human brain); synthesizing supercomputing (215KB,PDF), neuroscience, and nanotechnology (2.19MB,PDF); and attracting $53.5M of DARPA funding through four phases (read more about Phase 0, Phase 1, Phase 2, Phase 3) since 2008 has taken collaboration in the lab, across IBM Research and other business units, and with universities, government labs, and now, business partners and customers.
“In high school, I was the top student in math, physics, and chemistry across India’s second most-populous state. That feeling of individual accomplishment actually led me to subsequent failures—and taught me important lessons,” said Dharmendra, “Ego doesn’t achieve anything. Success is always the accumulation of many people’s work—and that’s how SyNAPSE has succeeded so far and has to succeed in the future.”
Computers have always been analogized with the brain. What makes SyNAPSE different? Today’s Von Neumann computer architecture, which has been the model for computers since 1946, is, at its core, a design for a fast calculator. It’s great for quickly crunching massive amounts of symbolic and structured data. But it cannot compete with the brain’s low power, small volume, and real-time performance for multi-modal sensory tasks, like recognizing the stripes of a tiger on the Savannah grasslands, or a friend’s face in a crowd.
Taking a radically different approach on computing by mimicking the brain’s anatomy and physiology (653KB,PDF), SyNAPSE’s neurosynaptic architecture is parallel, distributed and event-driven—and has its own end-to-end technical ecosystem. Like the brain, the architecture is innately scalable and has the potential to become a foundation for new classes of systems and services spanning mobile and portable devices, coprocessors, servers, supercomputers, and cloud.
“After my doctorate, IBM Research was the only place I wanted to work. No single thing persists except character. And IBM has the proven sustained character of making the world work better via continuous transformation and evolution.”
“So, when you create something at IBM, like SyNAPSE, it has the potential to endure,” Dharmendra said.
Dharmendra Modha in his own words
What does it mean to be named an IBM Fellow?
It significantly enhances our team’s ability to fully realize the SyNAPSE vision and its potential for IBM and all stakeholders.
I am grateful to my colleagues and collaborators; my managers; my mentors; DARPA; my alma maters, the University of California at San Diego, and the Indian Institute of Technology, Mumbai; IBM; IBM Research; The lab here in Almaden; and, of course, my family.
How do you come up with your best ideas?
Ideas can come from anywhere: team meetings, talking to my mentors, or during quiet mornings, evenings, weekends, and vacations. Even in sickness!
Ideas are just one part of the equation. The other part is to categorize, arrange, organize, and prioritize them as a prelude to decision and action. I use many creativity techniques such as six hats thinking, pro-cons, and SMUBA (seeing, mapping, understanding, believing, acting).
What do you consider the most important developments taking place today in Cognitive Computing?
Modern computing started with the U.S. social security administration for government; billing and inventory for business; and nuclear physics simulations for science. This lineage goes back to Blaise Pascal and early calculators. But with all the sensors around us creating data, we need (and now have!) a new computing architecture that’s designed and learns like the brain. More broadly speaking, as we come to the end of Moore’s law, application-specific hardware development will explode. However, this will necessitate software-hardware co-design—which has been a key principle underlying SyNAPSE.
What was the best piece of advice you ever received?
[IBM Senior Vice President and Director of Research] John Kelly advised me, “Don’t run from your problems, run to your problems.” I have used this mantra repeatedly to proactively seek out lemons on the horizon to turn them into lemonade.
What advice would you give to aspiring IBM Fellows?
Seek to solve problems of universal importance. To do that, collaborate, share credit, hoard blame, and never compromise on the vision.
What was your first job?
I started typing and then composing letters and documents for my father’s small business when I was 13, in seventh grade. He asked me because he could not speak English—but neither could I!
What do you enjoy doing away from the lab?
I enjoy spending time with my family, studying ancient literature, and doing yoga for exercise.
But I also collect limited edition Montblanc fountain pens. The Agatha Christie pen reminds me of the suspense and drama of the mysteries that I’m trying to solve. The Da Vinci pen reminds me of what man can achieve.
What are you reading right now?
Right now, Shakespeare’s Much Ado About Nothing, but I’ve also recently read ThinkerToys: A Handbook of Creative-Thinking Techniques by Michael Michalko, and Mistakes Were Made (But Not by Me) by Carol Tavris and Elliot Aronson.
What’s on your iPod right now?
I enjoy Indian classical music. Right now, I’m listening to Susheela Raman; she beautifully combines Eastern and Western music.
How would you like to be remembered?
I want to be remembered as someone who looked for the essence of everything, and possibilities everywhere to create something enduring … and as someone who, in collaboration with a stellar group of people, opened up radically new brain-inspired computing architectures, technology, systems, and ecosystems.