Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Hybrid computing using a neural network with dynamic external memory

nature.com

Post reply on HN
Print view View original

11–20 of 40 posts

Re: Hybrid computing using a neural network with dynamic external memory

#11

Post by modeless »

kylek wrote: post #8

I'm probably totally off base here (neural networks/AI is not my wheelhouse), but is having "memory" in neural networks a new thing? Isn't this just a different application of a more typical 'feedback loop' in the network?

The "memory" in a typical recurrent neural network is akin to a human's short term working memory. It only holds a few things and forgets old things quickly as new things come in. This new memory can hold a large number of things and stores them for an unlimited amount of time, more like a human's long term memory or a computer's RAM.

Re: Hybrid computing using a neural network with dynamic external memory

#12

Post by choxi »

kylek wrote: post #8

I'm probably totally off base here (neural networks/AI is not my wheelhouse), but is having "memory" in neural networks a new thing? Isn't this just a different application of a more typical 'feedback loop' in the network?

You're correct in a way, you can think of neural nets "remembering" the data set they're trained on. Recurrent neural nets even explicitly have a "feedback loop" like you're referring to that allows them to "remember" previous samples. An example of that is in natural language processing where you want to be able to remember the previous words in a sentence to interpret the current word.

Remembering the previous words in a sentence you're currently reading is more like short term memory though, and this paper is talking about long term memories stored as data structures outside of the neural net itself. This graphic from the DeepMind blog post might be helpful: https://i.imgur.com/KwXXCge.png.

The blog post from DeepMind is a bit more accessible than the Nature paper: https://deepmind.com/blog/differentiable-neural-computers/

Re: Hybrid computing using a neural network with dynamic external memory

#13

Post by AlexCoventry »

gallerdude wrote: post #3

Can someone explain what the full implications of this are? This seems really cool, but I can't really wrap my head around it. From what I can tell you can give the DNC simple inputs and it can derive complex answers.

It separates the concern of memorization from those of training and processing. In most current neural architectures, patterns in the training data are implicitly represented in the trained neural weights, and the net is implicitly forced to develop recall of past events by transmitting them from each time step to the next via neural net outputs.

The framework in this paper trains a neural net which interacts with a memory bank in a manner similar to a CPU. That means it can save and recall data on request, which could lead to more flexible architectures (you can give a trained net different data to recall) and easier training (since a memory-based architecture means the neural weights no longer have to learn the data along with the processing algorithm.)

Re: Hybrid computing using a neural network with dynamic external memory

#14

Post by AlexCoventry »

0xdeadbeefbabe wrote: post #5

> a DNC can complete a moving blocks puzzle in which changing goals are specified by sequences of symbols A neural network without memory can't do that or can't do it as well perhaps?

In fig. 5a, they compare its performance to that of an LSTM trained on the same problem, and it does seem to do much better.

Re: Hybrid computing using a neural network with dynamic external memory

#15

Post by triplefloat »

Very exciting extension of Neural Turing Machines. As a side note: Gated Graph Sequence Neural Networks (https://arxiv.org/abs/1511.05493) perform similarly or better on the bAbI tasks mentioned in the paper. The comparison to existing graph neural network models apparently didn't make it into the paper (sadly).

Re: Hybrid computing using a neural network with dynamic external memory

#18

Post by bbctol »

gallerdude wrote: post #16

Does this mean we could get way better versions of char-rnn?

This hopefully could replace current char-rnn with something very different. Char-rnn is a long short term memory system, where recurrence in the structure of the neural network allows short-term information to persist and inform future actions. This paper almost mimics the brain's separate long and short term memory structures, and could store long-term memory separately from its main activities until needed.
Post reply on HN