In the News
Elementary, My Dear Watson
Watson, IBM’s attempt to crack the market for artificial intelligence, is starting to be tested in the real world
Deep Learning robot takes 10 days to teach itself to grasp
Leave a human baby with some toys and it’ll quickly learn to pick them up. Now a robot with deep-learning capabilities has done the same thing.
Also in the news this week...
- Google made an investment in the German Research Centre for Artificial Intelligence (DFKI), a nonprofit institute with 450 scientists
- Boeing and Carnegie Mellon struck a $7.5 million deal to establish a new Aerospace Data Analytics Lab to work on projects that will apply AI and big data to improving the quality of Boeing’s aerospace activities
- Sony developed an artificial brain to estimate house prices and get an edge in the real estate market
- Youtube now uses deep learning to pick better video thumbnails
- Swiftkey launched its "neural network keyboard" and it's creepily good according to Gizmodo
- Berg Health claims it will release a cancer treating drug within 3 years, with a development time that was halved thanks to AI
- Android founder Andy Rubin looks beyond mobile and invests in artificial intelligence
Learning
The deception that lurks in our data-driven world
Interesting and thoughtful piece on our current approach to data and why we should rethink parts of this approach.
What to do with “small” data?
Many technology companies now have teams of smart data-scientists, versed in big-data infrastructure tools and machine learning algorithms, but every now and then, a data set with very few data points turns up and none of these algorithms seem to be working properly anymore. What the hell is happening? What can you do about it?
Top 5 arXiv Deep Learning Papers, Explained
The 5 papers are: "Training recurrent networks online without backtracking", "Semi-Supervised Learning with Ladder Network", "Towards Neural Network-based Reasoning", "Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks" and "LSTM: A Search Space Odyssey"
Unsupervised feature learning and Deep Learning tutorial
This tutorial will teach you the main ideas of Unsupervised Feature Learning and Deep Learning. By working through it, you will also get to implement several feature learning/deep learning algorithms, get to see them work for yourself, and learn how to apply/adapt these ideas to new problems.
Software tools & code
Speech and natural language processing resources
Great list of speech and natural language processing tools by Paul Dixon, a researcher based in Kyoto Japan.
Synaptic.js
This is a javascript architecture-free neural network library for node.js and the browser. Same principles as ConvnetJS, another JS library previously covered with a focus on convolutional nets.
Mindori: On-demand GPUs for Neural Networks
[Beta will launch in November] Service to help you spin up and down Titan Xs, with managed job queues and GPU drivers at a per-minute rate.
Some thoughts
Lessons learned from working at Continuum
Lessons Learned from Aaron Meurer who spent 2 years at Continuum primarily working on the Anaconda distribution and its open source package manager, conda.
Five principles for applying data science for social good
Every week, a data or technology company declares that it wants to “use data for good”. Beyond the good intentions, how can they and how could you actually achieve that?
About
This newsletter is a weekly collection of AI news and resources.
If you find this newsletter worthwhile, please forward to your friends and colleagues, or share on your favorite network!
Share on Twitter · Share on Linkedin · Share on Google+
As a thank you, next week's web and email issue will link to the top Twitter accounts that shared this letter.
Suggestions or comments are more than welcome, just reply to this email. Thanks!