Google has been given access to huge swatches of confidential patient information in the UK, raising fears yet again over how NHS managers view and handle data under their control. In an agreement uncovered by the New Scientist, Google and its DeepMind artificial intelligence wing have been granted access to current and historic patient data at three London hospitals run by the Royal Free NHS Trust, covering 1.6 million individuals. That would include any chronic illness people may be suffering from and the circumstances over why they were admitted – for example, if they have suffered a drug overdose. The agreement provides Google with access to data going back five years and is far more expansive than expected. Google and DeepMind previously said they were working with the NHS on a product called "Streams" that would "present timely information that helps nurses and doctors detect cases of acute kidney injury." The agreement however provides access to all patient data, covering issues far beyond just kidney functioning.

I recently gave a talk about recommender systems at the Data Science Sydney meetup (the slides are available here). This post roughly follows the outline of the talk, expanding on some of the key points in non-slide form (i.e., complete sentences and paragraphs!). The first few sections give a broad overview of the field and the common recommendation paradigms, while the final part is dedicated to debunking five common myths about recommender systems. The key reason why many people seem to care about recommender systems is money. For companies such as Amazon, Netflix, and Spotify, recommender systems drive significant engagement and revenue. But this is the more cynical view of things.

Our predictions are also indicated alongside draw history: successful predictions are marked with a tick. Failures (nothing's perfect) are marked with a cross. BigDataGame is powered by WSO2 Machine Learner, an open source machine learning solution that generates models for predictive data analysis. For more information on how BigDataGame was built, click here.

I've decided that I'm going to get really good at using R. What is R, you might ask? R is a programming language for originally created for statistical analysis, and my interest in it means I've apparently come to terms with the fact that I'm the type of person that jocks gave wedgies and swirlies to in high school. Why am I interested in R? Well, R is a programming language that will enable machine learning. I'm not talking about the robots-flipping-pancakes type of machine learning which -- while cool -- is not what comes to mind when thinking about statistics. Rather, the type of machine learning I'm talking about is where a user programs a computer to learn without being explicitly programmed to do so.

Nowadays, companies such as Google and IBM are doing a great service to all of us by showing what can be achieved by using Artificial Intelligence (AI). For instance, the results achieved by AlphaGo and Watson are outstanding and truly inspiring (the least one can say). But the fact that only companies and labs have access to this technology can represent a threat. First of all, we cannot know how money driven companies are going to use this novel technology. Second, this monopoly slows down technology adoption.

Earlier this week, OpenAI, the nonprofit research group with billion-dollar backing from Elon Musk and other tech luminaries, launched its first program. It's called OpenAI Gym, and it's meant to be used as a benchmarking tool for artificial intelligence programs. Musk once said he thought truly artificial intelligent agents could be more harmful to the human race than nuclear weapons. When OpenAI was launched in December, its stated goal was to "advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return." Which sounds a lot like an AI version of Google's long-held mantra: "Don't be evil."

My secretary, Amy Ingram (AI – geddit?), is designed to help people organise their diaries and set up meetings. That's all she can do. But she goes about it in a freakishly human-like way. In fact, after using her for a couple of weeks, various contacts of mine – after communicating with her over email – said they had no idea she was not a real person. For her to do her job, you need to give her access to your electronic diary, set a few preferences (your three favourite coffee shops to meet in, for example), and copy her into any emails about meetings you want to set up.

This has been the year of the AI machine, and it's been a rapid change. Artificial intelligence has suddenly begun to match and even outperform humans in tasks where we've have always held the upper hand--face recognition, object recognition, language understanding and so on. And yet there are plenty of complex tasks in which AI machines still trail humans. These range from simple housework such as ironing to more advanced tasks such as driving. The reason for the slow progress in these areas is not that intelligent machines can't do these tasks.

A visitor to Mobile World Congress in Barcelona tries on a virtual-reality headset, one of a coming wave of VR devices. SAN FRANCISCO – I love watching people experience virtual reality for the first time. The cumbersome headsets exaggerate their movements as they scan left to right, and then nod up and down. And every time they look in another direction and spot something new, you can almost feel their amazement through the goggles. Lately, I feel as though I've put on a VR headset and never took it off. Because every time I turn, it seems, I spy something truly inspirational.

There are two fundamental milestones I'd say. The first one is Fisher's Linear Discriminant [1], later generalized by Rao [2] to what we know as Linear Discriminant Analysis (LDA). Essentially, LDA is a linear transformation (or projection) technique, which is mainly used for dimensionality reduction (i.e., the objective is to find the k-dimensional feature subspace that -- linearly -- separates the samples from different classes best. Given the objective to maximize class separability, projecting the 2D dataset below onto "x-axis component," would be a better choice than the "y-axis component." Keep in mind though that LDA is a projection technique; the feature axes of your new feature subspace are (almost certainly) different from your original axes.