Convolutional Nets and CIFAR-10: An Interview with Yann LeCun

Date: December 24, 2014 Published by

In this blog series we will interview three contestants and also a founding father of convolutional nets: Yann LeCun. Many contestants used convolutional nets to tackle this competition. Who remembers that the first learning-based face detector that actually worked was a convolutional net (back in 1993, eight years before Viola-Jones)? Yes, deep nets need to be deep. One thing I’m excited about is the idea... View Article

Read More

Google’s Secretive DeepMind Startup Unveils a “Neural Turing Machine”

Date: December 23, 2014 Published by

They compare the performance of their Neural Turing Machine with a conventional neural network. Once again, the Neural Turing Machine significantly outperforms a conventional neural network. It turns out that the neural Turing machine learns to copy sequences of lengths up to 20 more or less perfectly. The DeepMind team go on to test the Neural Turing Machine on other tasks. The Neural Turing Machine... View Article

Read More

Mathematicians Make a Major Discovery About Prime Numbers

Date: December 23, 2014 Published by

A large prime gap is the same thing as a long list of non-prime, or “composite,” numbers between two prime numbers. The list then becomes 101! + 2, 101! + 3, 101! + 4, … , 101! + 101. Both teams did this by exploiting recent results — different ones in each case — about patterns in the spacing of prime numbers. The new work... View Article

Read More

How a Math Genius Hacked OkCupid to Find True Love

Date: December 20, 2014 Published by

Members answer droves of multiple-choice survey questions on everything from politics, religion, and family to love, sex, and smartphones. Chris McKinlay used Python scripts to riffle through hundreds of OkCupid survey questions. OkCupid’s algorithms use only the questions that both potential matches decide to answer, and the match questions McKinlay had chosen—more or less at random—had proven unpopular. OkCupid was founded by Harvard math majors... View Article

Read More

How Google Retooled Android With Help From Your Brain

Date: December 19, 2014 Published by

Last year Google researchers, including Dean, built a neural network program that taught itself to identify cats on YouTube. In November, a University of Toronto team used neural networks to predict how drug molecules might behave in the real world. It installed a voice recognition system based on what’s called a neural network — a computerized learning system that behaves much like the human brain.... View Article

Read More

Obama Goes Public With Brain-Mapping Plan

Date: December 19, 2014 Published by

Private foundations, including the Allen Brain Institute for Brain Science and Howard Hughes Medical Institute, will kick in tens of millions more. Some have argued that this approach to mapping brain activity is misguided and unlikely to yield important insights. The project has inevitably drawn comparisons — by advocates and detractors alike — to the Human Genome Project. The BRAIN initiative was preceded by the... View Article

Read More

Baidu Announces Breakthrough In Speech Recognition, Claiming To Top Google And Apple

Date: December 18, 2014 Published by

In noisy backgrounds, Ng said, tests showed that Deep Speech outperformed several speech systems–the Google Speech API, wit.ai, Microsoft’s Bing Speech, and Apple Dictation–by over 10% in terms of word error rates. Today, Baidu, often called China’s Google Google, unveiled the first results of what the former Google researcher, Stanford professor and Coursera cofounder had in mind. Baidu’s Deep Speech beat other methods such as... View Article

Read More

Getting Value From Your Data Scientists

Date: December 18, 2014 Published by

About one-third described themselves as data scientists; the rest identified themselves as analysts. We found substantial differences between the two. Similarly, 96% of data scientists said their analyses are used to make key decisions, versus 77% of quantitative analysts. For example, data scientists were far more likely to view their work as critical to better business outcomes. Today’s data scientists are often singled out as... View Article

Read More

‘Chinese Google’ Unveils Visual Search Engine Powered by Fake Brains

Date: December 18, 2014 Published by

The company has already deployed deep-learning algorithms for optical character, face, and voice recognition, online advertising and web search. But Google is Google, and mobile image search can be more engineering-intensive. Chinese search giant Baidu has served up its first ever visual search engine, which allows users to finally query the web using only images as input instead of keywords. Yu’s team is using Nvidia... View Article

Read More

How Skype Used AI to Build Its Amazing New Language Translator

Date: December 18, 2014 Published by

It’s called the Skype Translator—a Skype add-on that listens to the English words you speak into Microsoft’s internet phone-calling software and translates them into Spanish, or vice versa. Neural networks have been around since the 1980s, but they’re experiencing a renaissance. Very soon now, a select group of Skype beta testers will have a new Microsoft technology that seems borrowed from the world of Star... View Article

Read More