Category
24 posts
In the last blog we looked at the phenomenon of “jagged intelligence” — the paradox that AIs are simultaneously superhuman and stupid in adjacent skills. We identified a lack of continual, fast learning and embodiment as possible causes of this weirdness, and noted that embodiment is also a “learning problem”, because current ML models can’t learn from a single thread of…
In continual learning, the neural network learns from a stream of data, acquiring new knowledge incrementally. It’s not possible to assume an i.i.d. stationary dataset available in one batch. Catastrophic forgetting of previous knowledge is a well known challenge. A wide variety of approaches fall broadly into 3 categories [6]: Regularization-based methods, Parameter isolation methods…
We recently presented 2 papers at the International Joint Conference on Neural Networks, IJCNN. The first one is about one-shot learning for the long term (with an artificial hippocampal algorithm), blog here. In this blog article, we are excited to share the other paper – Learning distant cause and effect using only local and immediate credit assignment…
We’ve just undertaken a review and refresh of our research roadmap! The topics and approach we will take in the coming year are all laid out in a new page: Research Roadmap Our primary topics for the coming year include: Continual Few-Shot Learning (CFSL) via our Episodic memory system Using reinforcement learning to control a…
We’re very happy to report that we recently published a preprint on AHA, an ‘Artificial Hippocampal Algorithm’ for Episodic Machine Learning. It’s the culmination of a multi-year research project and is a starting point for the next wave of developments. This article describes the motivation for developing AHA and a brief description of the research.…
One of our key projects is a memory system that can learn to associate distant cause & effect while only using local, immediate & unsupervised credit assignment. Our approach is called RSM – Recurrent Sparse Memory. We recently uploaded a preprint describing RSM. This is the first of several blog posts about the experiments we…
We’ve uploaded a new paper to arXiv presenting our algorithm for biologically-plausible learning of distant cause & effect using only local and immediate credit assignment. This is a big step for us – it ticks almost all our requirements for a general purpose representation. The training regime is unsupervised & continuous. Local and immediate credit…
Adaptive optimization methods, such as Adam and Adagrad, maintain some statistics over time about the variables and gradients (e.g. moments) which affect the learning rate. These statistics won’t be very accurate when working with sparse tensors, where most of its elements are zero or near zero. We investigated the effects of using Adam with varying…
We recently talked about Capsules networks and equivariances. NB: If you’re not familiar with Capsules networks, read this first. Our primary objective with Capsules networks is to exploit their enhanced generalization abilities. However, what we’ve found instead raises new questions about how generalization can be measured and whether Capsules networks are actually good at it. PredCaps…
It’s such a joy to be able to test an idea, go straight to the idea without wrestling with the tools. We recently developed an experimental setup which, so far, looks like it will do just that. I’m excited about it and hope it can help you too, so here it is. We’ll go through the why we created another framework, and how each module in the experiment setup works.
We are exploring the nature of equivariance, a concept that is now closely associated with the capsules network architecture (see key papers Sabour et al, and Hinton et al). Machine learning representations that capture equivariance must learn the way that patterns in the input vary together, in addition to statistical clusters in the input (that a typical…
This is the second part of our comparison between convolutional competitive learning and convolutional or fully-connected sparse autoencoders. To understand our motivation for this comparison, have a look at the first article. We decided to compare two specific algorithms that tick most of the features we require: K-Sparse autoencoders, and Growing-Neural-Gas-with-Utility (GNG-U) (Fritzke1997). The latter…
Competitive learning is a branch of unsupervised learning that was popular a long, long time ago in the 1990s. Older readers may remember – the days before widespread use of GSM mobile phones and before Google won the search engine wars! Although competitive learning is now rarely used, it is worth understanding the broader range of…
We’ve just uploaded a spin-off research paper to arXiv titled “Sparse Unsupervised Capsules Generalize Better”. So what’s it all about? Capsules Networks You may have heard of Capsules Networks already – if not, have a read of one of these blog articles (here, here, here, or here (EM routing)), watch this video, or consult one of the…
The dataset is an integral part of an ML engineer’s toolkit. We recently compiled useful information about a range of these well known datasets. It’s all in one place, and hopefully useful to others as well.
There are plenty of established machine learning frameworks out there, and new frameworks are popping up frequently to address specific niches. We were interested in examining if one of these frameworks fits in our workflow. I surveyed the most popular frameworks, and aim to provide a helpful comparative analysis.
SVHN is relatively new and popular dataset, a natural next step to MNIST and complement to other popular computer vision datasets. This is an overview of the common preprocessing techniques used and the best performance benchmarks, as well as a look at the state-of-the-art neural network architectures used.
Releasing a set of tools for converting the Street View House Numbers (SVHN) dataset into images with additional preprocessing options such as grayscaling.
We’re very excited to launch AGI Experimental Framework, AGIEF, our open source framework. We first introduced it a while back, at the end of 2015 here, and it has certainly come a long way. AGIEF was created to make running rigorous AI experiments convenient, reproducible and scalable. The goals are: Repeatability: ability to save/load, stop/start…
In our most recent post we discussed the current set of experiments that we are conducting, using the MNIST dataset. We’ve also been looking at the NIST dataset which is similar, but extends to handwritten letters (as well as digits). These are extremely popular datasets and freely available, so make a great choice for testing…
Typical results from our experiments: Some active cells in layer 3 of a 3 layer network, transformed back into the input pixels they represent. The red pixels are positive weights and the blue pixels are negative weights; absence of colour indicates neutral weighting (ambiguity). The white pixels are the input stimulus that produced the selected…
Erik Laukien is back with a demo of Sparse, Distributed Representation with Reinforcement Learning. This topic is of intense interest to us, although the problem is quite a simple one. SDRs are a natural fit with Reinforcement Learning because bits jointly represent a state. If you associate each bit-pattern with a reward value, it is…
By Gideon Kowadlo and David Rawlinson Introduction We’ve been building and testing AGI algorithms for the last few years. As the systems become more complex, we have found it ever more difficult to run meaningful experiments. To summarise, the main challenges are: testing a version of the algorithm repeatedly and over some range of parameters…
This might be our new test suite for the algorithms! http://www.arcadelearningenvironment.org/ “The Arcade Learning Environment (ALE) is a simple object-oriented framework that allows researchers and hobbyists to develop AI agents for Atari 2600 games. It is built on top of the Atari 2600 emulator Stella and separates the details of emulation from agent design.” Why…