Google’s Secretive DeepMind Startup Unveils a “Neural Turing Machine”

Google’s Secretive DeepMind Startup Unveils a “Neural Turing Machine”

From the post:

One of the great challenges of neuroscience is to understand the short-term working memory in the human brain. At the same time, computer scientists would dearly love to reproduce the same kind of memory in silico.

Today, Google’s secretive DeepMind startup, which it bought for $400 million earlier this year, unveils a prototype computer that attempts to mimic some of the properties of the human brain’s short-term working memory. The new computer is a type of neural network that has been adapted to work with an external memory. The result is a computer that learns as it stores memories and can later retrieve them to perform logical tasks beyond those it has been trained to do.

Of particular interest to topic mappers and folks looking for realistic semantic solutions for big data. In particular the concept of “recoding,” which is how the human brain collapses multiple chunks of data into one chunk for easier access/processing.

It sounds close to referential transparency to me but where the transparency is optional. That is you don’t have to look unless you need the details.

The full article will fully repay the time to read it and then some:

Neural Turing Machines by Alex Graves, Greg Wayne, Ivo Danihelka.


We extend the capabilities of neural networks by coupling them to external memory resources, which they can interact with by attentional processes. The combined system is analogous to a Turing Machine or Von Neumann architecture but is differentiable end-to-end, allowing it to be efficiently trained with gradient descent. Preliminary results demonstrate that Neural Turing Machines can infer simple algorithms such as copying, sorting, and associative recall from input and output examples.

The paper was revised on 10 December 2014 so if you read an earlier version, you may want to read it again. Whether Google cracks this aspect of the problem of intelligence or not, it sounds like an intriguing technique with applications in topic map/semantic processing.

Comments are closed.