Algorithm Development Quotes

To write a kernel without a data structure and have it be as consistent and graceful as UNIX would have been a much, much harder challenge.

Before Google, I spent the summer building a program that would look at what websites you would go to and what websites other people would go to - and built a collaborative filtering program that helped you find related sites to look at.

Back in the really early days, the men went out hunting, the women stayed home with the kids, and would hold the kid in one arm against the heart, so that's the left, and with the right arm they would throw. And it turns out you cannot make that calculation in real time. You have to have an algorithm set up. So these brain mechanisms evolved in order to do that, and when they evolved, the thing is that where there is a useful capability it often adapts to places it wasn't evolved for.

We have a good sense of the pulse of the industry. We know what's going on out there. We know that deep learning adoption is broad. It's going into production at scale.

Formula creation is magnitudes harder for computer algorithms than actually executing within a formula. But once the genre is created and the formula is known, then the computer can do the repetitive task of executing within the genre.

The demise of Google Reader, if logical, is a reminder of how far we've come from the cuddly old 'I'm Feeling Lucky' Google days, in which there was a foreseeably-astonishing delight in the way Google's evolving design tricks anticipated what users would like.

Indeed, nobody is close to developing a general formula that predicts the number flips required for any given number of pancakes.

After all, if you had the complete decision process, you could run it as an AI, and I'd be coding it up right now.

You will be able to program a robot to follow a track on the ground and manipulate a hand. You can also write little programs that will give the robots goals.

This is where the world is going: direct access from anywhere to any type of data, whether it's a small piece of data or a small answer but a long algorithm to create that answer. The user doesn't care about this.

Beyond helping other people build AI systems with Deeplearning.ai, I also hope to build some AI systems myself!

When I was working on my Ph.D., I developed a computer algorithm to look for rapid changes in populations' DNA. Our DNA changes constantly over generations, but if certain changes spread through a population more quickly than others, they are probably the beneficial results of natural selection. This is the protection we give ourselves to survive.

Data science is the combination of analytics and the development of new algorithms.

When Spotify launched in the U.S. in 2011, it relied on simple usage-based algorithms to connect users and music, a process known as 'collaborative filtering.' These algorithms were more often annoying than useful.

I just thought making machines intelligent was the coolest thing you could do. I had a summer internship in AI in high school, writing neural networks at National University of Singapore - early versions of deep learning algorithms. I thought it was amazing you could write software that would learn by itself and make predictions.

It took me 1057 pages to describe the hundreds of mathematical equations, algorithms and programming techniques that I invented and used.

In their field they [mathematicians] do what we ought to be doing in ours. Therein lies the significant lesson ... of their existence. They are an analogy for the intellectual of the future.

In deep learning, the algorithms we use now are versions of the algorithms we were developing in the 1980s, the 1990s. People were very optimistic about them, but it turns out they didn't work too well.

I took a computer-science course to fill a prerequisite at Stanford, and I realized that every day was a new problem, and every day you got to think about how to solve something new, how to reason through something new, how to develop an algorithm to solve for something you hadn't worked on before.

One of the things that Baidu did well early on was to create an internal platform for deep learning. What that did was enable engineers all across the company, including people who were not AI researchers, to leverage deep learning in all sorts of creative ways - applications that an AI researcher like me never would have thought of.

One of the things Baidu did well early on was to create an internal platform that made it possible for any engineer to apply deep learning to whatever application they wanted to, including applications that AI researchers like me would never have thought of.

AIs are only as good as the data they are trained on. And while many of the tech giants working on AI, like Google and Facebook, have open-sourced some of their algorithms, they hold back most of their data.

When The Daily Muse initially wanted to launch a job board, our first ideas were insanely (and needlessly) complex. We wanted to integrate with social networks, gather rich personal data to build predictive algorithms, and put together numerous cool visualization tools before launching out to the world. We were just sure users would love it!

As our understanding of fraud evolves, we might one day be able to develop predictive algorithms that could identify would-be con artists based on patterns of behavior.

I thought of the idea of Summly in March or April 2011. I was 15 years old and I was revising for some kind of history exam. The problem was I was trying to find information that was useful to me. When you type into Google an esoteric term, you get quite a lot of stuff that's not relevant.