From Google's purchase of Deepmind to Apple's purchase of VocalIQ, big deals in UK's intelligence technologies are booming. However, commercial innovations are only part of the story. Join Digital Catapult to meet SMEs, academics and investors working in this space. It's a fantastic opportunity to find out what the UK's leading academic and research organisations and businesses are doing and attending, you'll also gain exclusive insights into market opportunities, challenges and exciting advances.

Bayesian methods for machine learning have been widely investigated, yielding principled methods for incorporating prior information into inference algorithms. In this survey, we provide an in-depth review of the role of Bayesian methods for the reinforcement learning (RL) paradigm. The major incentives for incorporating Bayesian reasoning in RL are: 1) it provides an elegant approach to action-selection (exploration/exploitation) as a function of the uncertainty in learning; and 2) it provides a machinery to incorporate prior knowledge into the algorithms. We first discuss models and methods for Bayesian inference in the simple single-step Bandit model. We then review the extensive recent literature on Bayesian methods for model-based RL, where prior information can be expressed on the parameters of the Markov model. We also present Bayesian methods for model-free RL, where priors are expressed over the value function or policy class. The objective of the paper is to provide a comprehensive survey on Bayesian RL algorithms and their theoretical and empirical properties.

We propose Very Simple Classifier (VSC) a novel method designed to incorporate the concepts of subsampling and locality in the definition of features to be used as the input of a perceptron. The rationale is that locality theoretically guarantees a bound on the generalization error. Each feature in VSC is a max-margin classifier built on randomly-selected pairs of samples. The locality in VSC is achieved by multiplying the value of the feature by a confidence measure that can be characterized in terms of the Chebichev inequality. The output of the layer is then fed in a output layer of neurons. The weights of the output layer are then determined by a regularized pseudoinverse. Extensive comparison of VSC against 9 competitors in the task of binary classification is carried out. Results on 22 benchmark datasets with fixed parameters show that VSC is competitive with the Multi Layer Perceptron (MLP) and outperforms the other competitors. An exploration of the parameter space shows VSC can outperform MLP.

Hamiltonian Monte Carlo (HMC) is a popular Markov chain Monte Carlo (MCMC) algorithm that generates proposals for a Metropolis-Hastings algorithm by simulating the dynamics of a Hamiltonian system. However, HMC is sensitive to large time discretizations and performs poorly if there is a mismatch between the spatial geometry of the target distribution and the scales of the momentum distribution. In particular the mass matrix of HMC is hard to tune well. In order to alleviate these problems we propose relativistic Hamiltonian Monte Carlo, a version of HMC based on relativistic dynamics that introduce a maximum velocity on particles. We also derive stochastic gradient versions of the algorithm and show that the resulting algorithms bear interesting relationships to gradient clipping, RMSprop, Adagrad and Adam, popular optimisation methods in deep learning. Based on this, we develop relativistic stochastic gradient descent by taking the zero-temperature limit of relativistic stochastic gradient Hamiltonian Monte Carlo. In experiments we show that the relativistic algorithms perform better than classical Newtonian variants and Adam.

Thanks to the modern gaming industry, we can now spend our evenings wandering around photorealistic game worlds, like the post-apocalyptic Boston of Fallout 4 or Grand Theft Auto V's Los Santos, instead of doing things like "seeing people" and "engaging in human interaction of any kind." Games these days are so realistic, in fact, that artificial intelligence researchers are using them to teach computers how to recognize objects in real life. Not only that, but commercial video games could kick artificial intelligence research into high gear by dramatically lessening the time and money required to train AI. "If you go back to the original Doom, the walls all look exactly the same and it's very easy to predict what a wall looks like, given that data," said Mark Schmidt, a computer science professor at the University of British Columbia (UBC). "But if you go into the real world, where every wall looks different, it might not work anymore." Schmidt works with machine learning, a technique that allows computers to "train" on a large set of labelled data--photographs of streets, for example--so that when let loose in the real world, they can recognize, or "predict," what they're looking at.

Sam Kronick has a bunch of rocks arrayed in front of him on a raised desk in his Oakland studio. He's an artist and his plan is to sketch the rocks, but not with pen and paper. He and his artistic partner Tara Shi are going to do a 3D scan of them so that an artificial intelligence program can map their contours, learn to recognize rocks and then start generating its own craggy depictions. The project is deceptively simple: trying to get artificial intelligence to make nature art. Kronick and Shi are using a neural net, a computer program loosely modeled on biological neural systems like the human brain.

Capital One Europe CIO Rob Harding is leading a huge transformation at the company's UK business which involves becoming more digital, agile and self-sufficient, as he told CIO UK recently when he described his six-point master plan of fusing financial services with technology. As part of this transition, Harding has been keeping his eye on the biggest IT trends. In fact, he keeps a graph that he updates periodically of all the emerging technologies and buzzwords. The x-axis of the graph is Harding's opinion on the importance to CapitalOne UK's business strategy and the y-axis is the firm's current fluency with the technology. Below we run down his thoughts on several of these trends.

Tech columnist Jennifer Jolly takes a spin in a self-driving Ford Fusion and gets the scoop on how the technology works. Mercedes-Benz, whose engineers have been working on self-driving car technology, is eager to increase the size of its engineering team both in Silicon Valley and in Germany. SAN FRANCISCO -- So you say you want join the automotive revolution? Over the past few years, only elite roboticists have been positioned to heed the self-driving car's call to action. Armed with degrees from places such as Carnegie Mellon University and experience at institutions such as NASA, these tech whizzes have been highly sought after by technology and automotive companies looking to build the future.

Mercedes-Benz, whose engineers have been working on self-driving car technology, is eager to increase the size of its engineering team both in Silicon Valley and in Germany. SAN FRANCISCO - So you say you want join the automotive revolution? Over the past few years, only elite roboticists have been positioned to heed the self-driving car's call to action. Armed with degrees from places such as Carnegie Mellon University and experience at institutions such as NASA, these tech titans have been highly sought after by technology and automotive companies looking to build the future. But now massive open online course pioneer Udacity has a proposition: Give the Web-based education outfit 36 weeks and 2,400, and they'll turn graduates onto jobs at autonomous-car partner companies Mercedes-Benz, Didi Chuxing, Nvidia and Otto.

This turns out to have been an awkward thing for Wells Fargo Chief Executive Officer John Stumpf to have said about Carrie Tolstedt, Wells Fargo's head of Community Banking, when she announced her retirement in July: "A trusted colleague and dear friend, Carrie Tolstedt has been one of our most valuable Wells Fargo leaders, a standard-bearer of our culture, a champion for our customers, and a role model for responsible, principled and inclusive leadership," said John Stumpf, Wells Fargo's chairman and chief executive officer. It turns out that Wells Fargo's community banking culture involved creating millions of fake accounts for customers to satisfy the bank's frenzy for cross-selling products and services. And Tolstedt now gets to bear the standard for that culture, as the Senate investigates, Fortune reports that "she will be walking away with 124.6 million in stock, options, and restricted Wells Fargo shares," and shareholders have called for her to be held responsible for the fake accounts by clawing back her pay: Another investor said: "If this person presided over this, why no accountability? We have share-based pay so that it can be clawed back when people have been earning bonuses under false pretences, and if fraudulently opening client accounts isn't false pretences, then I don't know what is." Wells Fargo's cross-selling scandal is so odd because it is both at the absolute core of the bank's business, and also curiously irrelevant.