Email of the day on technological innovation
Comment of the Day

January 09 2014

Commentary by Eoin Treacy

Email of the day on technological innovation

Saw your clip about 3D scanners. Thanks for this. The data issue is one that I have not thought of.

Then I got this link about "deep learning" which may be of further interest to you.

Best wishes

Eoin Treacy's view

Thank you for your well wishes and this interesting article. While one might be bemused at how many videos of cats there are on the internet, there is no denying the digital age has ushered in an acceleration in the pace of innovation. The fact that discoveries can then be communicated to a wide audience virtually instantly further boosts the pace of data creation. Here is a section from the article you mentioned:

When Google adopted deep-learning-based speech recognition in its Android smartphone operating system, it achieved a 25% reduction in word errors. ?¡ãThat's the kind of drop you expect to take ten years to achieve,?¡À says Hinton "a reflection of just how difficult it has been to make progress in this area. "That's like ten breakthroughs all together."

Meanwhile, Ng had convinced Google to let him use its data and computers on what became Google Brain. The project's ability to spot cats was a compelling (but not, on its own, commercially viable) demonstration of unsupervised learning a the most difficult learning task, because the input comes without any explanatory information such as names, titles or categories. But Ng soon became troubled that few researchers outside Google had the tools to work on deep learning. "After many of my talks,?¡À he says, "depressed graduate students would come up to me and say: 'I don't have 1,000 computers lying around, can I even research this?"

So back at Stanford, Ng started developing bigger, cheaper deep-learning networks using graphics processing units (GPUs) the super-fast chips developed for home-computer gaming. Others were doing the same. ?¡ãFor about US$100,000 in hardware, we can build an 11-billion-connection network, with 64 GPUs,?¡À says Ng. 

The advances in neural networking and deep-learning have been made possible by the data rich environment in which we now live. Just as car ownership really took off with the building of highways, innovations in how we crunch, interpret and learn from data sets is likely to grow exponentially as the amount of data we have to work with continues to expand. A number sites already highlight how investment vehicles are trending on social media or how many times a keyword is searched for on Google. These probably represent the thin edge of the wedge in how data can be interpreted to discover additional information about companies and perceptions. Nevertheless, price action will always tell us what people are actually doing with their money. 

Moving on, the UN released its annual World Intellectual Property Indicators report in December and charts beginning on page 48 highlight the fact that the pace of patent registration has increased from 1 million a year in 1995 to 2.5 million today. Additionally, while the USA continues to successfully attract more inventors than any other country, China, India and South Korea are now important centres of patent registration. This would suggest that while the USA still has a considerable technological lead, the prospects for Asian countries to successfully innovate are increasingly bright. 

Back to top

You need to be logged in to comment.

New members registration