Tag Archives: Machine Learning

Deep Learning Update

This TED talk by Jeremy Howard in Brussels, created in December 2014, reveals the staggering progress made in the field of deep learning:

Jeremy Howard’s TED Talk on Deep Learning

In this TED talk, he speaks about:

– Amazon and NetFlicks use machine learning to suggest products that you would like.
_ IBM’s Watson beat the two world champions in Jeopardy
– Google’s car has now driven without a driver over a million miles without an accident.
– Jeffrey Hinton beat all the others, in just two weeks, to identify new drugs.
– Deep Learning learned how to recognize a wide variety of German Street signs.

He demonstrates:

– that computers can see …. image recognition application where 1.5 million pictures of cars are classified, where a human help the machine learn by “training” it to recognize “front”, “back” “angle” etc. He says that there are 16,000 dimensions to the analysis. He asks: can a pathologist look for areas of mitosis? Can a radiologist ….? A second image Stanford application where a computer can look at an image and describe in text, with some success, what is the image about. Humans asked about the text preferred the computer description 25% of the time – he predicts it will pass human performance in less than a year.
– computers can understand …. showing how a Stanford-based approach can read a sentence and understand the sentiment.
– computers can search images …. and match them to text …. he points out that this breakthrough is just in the last few months. The approach by Google searches text tags of the image (and thus is not this)
– that computers can listen ….voice recognition application where an English speaker can have his voice (not another voice) translated real-time into Chinese.
– computers can write ….

He speaks about the exponential growth in understanding which is underway. He believes that in the next five years, machine Learning performance will exceed human learning performance.

He does not believe better education will help. He thinks now is the time to begin adjusting our social structures to accommodate this new world.

He also speaks about applications:
– medical diagnostics through analysis of cancer tissue.

From Wikipedia:
Jeremy Howard
Jeremy Howard (born 1973) is an Australian data scientist and entrepreneur.[3] He is the CEO and Founder at Enlitic, an advanced machine learning company in San Francisco, California. Previously, Howard was the President and Chief Scientist at Kaggle, a community and competition platform of over 200,000 data scientists. Howard is the youngest faculty member at Singularity University, where he teaches data science. He is also a Young Global Leader with the World Economic Forum, and spoke at the World Economic Forum Annual Meeting 2014 on “Jobs For The Machines.”[4] Howard advised Khosla Ventures as their Data Strategist, identifying the biggest opportunities for investing in data driven startups and mentoring their portfolio companies to build data-driven businesses. Howard was the founding CEO of two successful Australian startups, FastMail and Optimal Decisions Group. Before that, he spent eight years in management consulting, at McKinsey & Company and AT Kearney.

Kaggle
Howard first became involved with Kaggle, founded in April 2010,[8] after becoming the globally top-ranked participant in data science competitions in both 2010 and 2011. The competitions that Howard won involved tourism forecasting[1] and predicting the success of grant applications.[2] Howard then became the President and Chief Scientist of Kaggle.[9]

Enlitic
In August 2014, Howard founded Enlitic with the mission of leveraging recent advances in machine learning to make medical diagnostics and clinical decision support tools faster, more accurate, and more accessible. Enlitic uses state-of-the-art Deep Learning algorithms to diagnosis illness and disease.[13] Howard believes that today, machine learning algorithms are actually as good as or better than humans at many things that we think of as being uniquely human capabilities.[14] He projects that the application of deep learning will have the most significant impact on medicine out of any technology during this decade by effectively aggregating data.[15] On October 28, 2014, Howard announced Enlitic’s seed funding round.[16]