Get email delivery of the Cadence blog featured here
“Intuitions are not to be ignored, John. They represent data processed too fast for the conscious mind to comprehend.”
“… No human action is ever truly random. An advanced grasp of the mathematics of probability mapped onto a thorough apprehension of human psychology and the known dispositions of any given individual can reduce the number of variables considerably.”
― Mark Gatiss and Steven Moffat, in Sherlock Holmes, “The Six Thatchers”
I have always thought that intuition is just our monkeybrain’s way of interpreting so much data that even though we can’t put a word to why we intuitively know something, we often act on that intuition, with no explanation why. And then when we find out later that we were right, the first thing that people say is, “I alllllllways knew something was strange about that guy...” “Like what?” we might be asked. “I don’t know, but there was just this feeling I had…”
“Intuition” is the word that we assign to processing vast amounts of data for which we may not even have words.
At the recent Embedded Vision Summit, I found myself thinking about this a lot. Neural networks and machine learning are based on the human model. They were designed alongside our understanding of the way the human brain works.
Jitendra Malik, Professor of Electrical Engineering and computer science at UC Berkeley, also noted this in his keynote on the second day of the conference, “Deep Visual Understanding from Deep Learning”. In it, he compared the primate visual system with where we are now in the architectures of neural networks.
(Note: keep an eye on Breakfast Bytes coming up; I know that Paul will be writing about Jitendra’s keynote – which contained a lot of incredible research and revelations – and other fascinating sessions that he attended at the summit!)
While neural networks are powerful in discrimination and generalization, we have a long, long way to go for them to have an “intuition” of their own. I look at Jitendra’s slide again, and shake my head.
This is not to say that I despair. No one did at this summit. Everyone seemed optimistic and realistic about the challenges moving forward. As I mentioned in my previous two posts about the event, it was all about how we must address the issues facing this new technology, which is poised to either:
a) Explode in our faces, or
b) Explode under our backsides, propelling us to Mars.
Every single session I attended – with no exceptions – began with some unique way of saying that neural networks require a lot of processing power, and we haven’t figured out how to do it effectively enough. Cadence is doing its part, that’s for sure. I kept wanting to raise my hand and ask, seriously? Have you heard about the announcement of the Cadence Tensilica Vision C5 DSP? But I didn’t want to sound like a salesperson or shill, which I am absolutely not.
The second thing that was raised consistency through most of the sessions was the fact that we have a dearth of specialists in this field, which leads me to my second point.
Many of these presentations, at least in the Business Insights Track, presented the problem that we’re on the cusp of reinventing the world of computer engineers. One of them even gave some numbers to back this up: while there are about 2 million computer programmers on earth, there are only about 20,000 people involved in machine learning technology. Those of us who are here at the beginning have a great deal of influence on the way this technology grows. It is here that the structure – the groundwork – is being laid.
So why is it that I could probably count on one hand the number of women in each session that I attended?
Sure, attendance is up by 20% from last year, per the Embedded Vision Summit itself, but I am very interested in finding out the demographics of attendance. I am used to being in the minority at conferences and trade shows – that’s nothing new – but here it is particularly evident. What is it about this technology that appeals only to men? Speaking of intuition… we women do have it. I’m just saying, is all.
I see only one woman in this photo of about 35 people on the show floor. I'd guess that this male: female ratio is commensurate with the entire body of the attendees.
What’s up with that? I assume most people reading this blog are male, and I'm interested in your thoughts about it. Do you agree that this is kind of a problem? It is certainly remarkable, and I mean that in the literal sense. There is a striking imbalance in the field, even with the gains that women have made in the STEM fields in the last twenty years or so.
And why does a woman blogger have to point this out? I would so love to see a man say that this is important to look into, too.
I’m interested in your thoughts. And expect more on this from me.