4 ms·
A lot of the datasets behind the big image recognition and reinforcement learning (especially recently) have been open sourced: - https://research.googleblog
by choxi 10y ago
A lot of the datasets behind the big image recognition and reinforcement learning (especially recently) have been open sourced:
- https://research.googleblog.com/2016/09/announcing-youtube-8m-large-and-diverse.html?m=1 https://research.googleblog.com/2016/09/announcing-youtube-8...
- https://techcrunch.com/2016/10/05/udacity-open-sources-an-additional-183gb-of-driving-data/ https://techcrunch.com/2016/10/05/udacity-open-sources-an-ad...
- http://image-net.org/ http://image-net.org/
There are also more paid, private dataset services now:
- https://fantasydata.com/pricing/nfl-data-api.aspx https://fantasydata.com/pricing/nfl-data-api.aspx
- https://www.quandl.com/ https://www.quandl.com/
I could see there being a lot of opportunity for startups that democratize data, which would then unlock more startups to build novel ML architectures on top of.
- Chronic9q 10y agoI know for a fact (and you can check online too), that Google's non-public datasets are magnitudes larger in training size and far superior in learning signal (labels are very close to their task). Google has thousands of real robotic manipulator arms to train their reinforcement learning algorithms. Algorithms trained using the OpenAI gym are practically guaranteed to fail in the real world since OpenAI gym is (designed as) a perfect simulator. Don't even get me started on internal image and speech recognition datasets collected from YouTube, search indexing, and Hangouts/Duo calls. There is no way in hell you can compete with this. The closest you have is Facebook. But they've built an AI that has just learned how to play Go. Let alone defeat a world champion. Hell, Google is already working on beating humans at Starcraft.