What AI still can’t do by Brian Bergstein via @techreview

Humans aren’t very good at understanding causation either.

I agree with Brian that we need to change our current AI thinking and combine a lot more sources of information using neural networks in order to get much better results.

Google AI Blog:Predicting the Generalisation Gap in Deep Neural Networks by Yiding Jiang via @googleai

Here’s a description of a new technique that uses margin distributions to better predict a DNN’s generalization gap.

Seems a great idea to use what the Google AI team have made available in their Github is a great idea and should not be ignored. Links to a lot of sources are given thought the article.

Open sourcing wav2letter++, the fastest state-of-the-art speech system, and flashlight, an ML library going native by/via via @fbOpenSource

The Facebook AI Research (FAIR) Speech team is sharing the first fully convolutional speech recognition system. It uses convolutional neural networks (CNNs) for acoustic modeling and language modeling, and is reproducible. The team says that wav2letter++ is composed only of convolutional layers, which yields performance that’s competitive with recurrent architectures.

There are two articles linked of the landing page from the links in this post. This reads as a great achievement and looks very interesting.

Convolutional Neural Net in Tensorflow by Stephen Barter via @Medium

Here's a look at the fundamentals of convolutional neural nets and how you can create one yourself to classify handwritten digits.

This is a great guide and I think it is well worth a subscription to see what else the author has written on Medium - so much in this article to learn from.

How DeepMind's biggest AI project is fixing bad Android batteries by Matt Burgess via @WiredUK

Google's Android Pie operating system uses DeepMind's AI in a bid to improve your phone's battery life. But is it making any difference?

This sounds great and of course over time it will get even better.

Convolutional neural networks for language tasks by Garrett Hoffman via @OReillyMedia

Though they are typically applied to vision problems, convolution neural networks can be very effective for some language tasks.

This looks really useful and it has snippets of code for you too.

Deep Learning and Google Street View Can Predict Neighbourhood Politics from Parked Cars by @BotJunkie via @IEEESpectrum

It's likely that your car says something about you. The make and model, whether it's foreign or domestic, and how expensive it is can provide information about who owns it.

This is very interesting.  I think this is not true for everyone, but you can certainly make all of these assumptions.

An Introduction to different Types of Convolutions in Deep Learning by @gopietz via @Medium

Behind the “C” in “CNN”. If you’re not super-familiar with how the internals of image classifiers work, this is a useful intro.

I love this and the great diagrams are really slick and help understanding a great deal.

Deep Learning Key Terms, Explained by Matthew Mayo via @kdnuggets

Gain a beginner's perspective on artificial neural networks and deep learning with this set of 14 straight-to-the-point related key concept definitions.

A great list of terms that you need to read and learn from.

A Beginner’s Guide To Understanding Convolutional Neural Networks Parts 1 and 2 by Adit Deshpande via @kdnuggets

Interested in better understanding convolutional neural networks?

Here are Part 1 and Part 2 - both are 2 pages long

Wow - I read these both about 3 times each before I felt I understood it completely.

How Convolutional Neural Networks Work by Brandon Rohrer via @kdnuggets

Get an overview of what is going on inside convolutional neural networks, and what it is that makes them so effective.

This is very clear and easy to understand. It is a two page article.