An easy day trip from London, Bletchley Park reveals there's far more to Britain's wartime code breaking than Alan Turing's battle with the Nazi Enigma machine. There are plenty of cool nerdy places to visit in London, from the British Library to Greenwich, but if you've got a little more time on your hands it's worth the trip up to Bletchley Park to see the how the Brits turned the tide of the war by eavesdropping on Hitler's seemingly uncrackable wartime communications systems. Bletchley Park mansion is roughly an hour's drive north of London, just off the M1. Alternatively you can catch the London Midland train – the mansion is only a short walk from Bletchley Park station – or else the Virgin Train to Milton Keynes Central and take a quick train or bus trip back to Bletchley Park. Purchased by MI6 just before the outbreak of the Second World War, the mansion initially served as home to the Government Code and Cypher School.

Researchers from Switzerland have confirmed what most of us already know - drinking a single glass of beer can make people more sociable. The team from University Hospital in Basel tested 60 healthy people, with an equal number of men and women drinking alcoholic and non-alcoholic beer. They took part in a range of tasks, including a face recognition test, empathy test and sexual arousal test. The lead researcher said there had been little previous research in this area. Prof Matthias Liechti explained: "Although many people drink beer and know its effects through personal experience there is surprisingly little scientific data on its effects on the processing of emotional social information."

President Vladimir Putin's ruling party easily outdistanced its rivals in parliamentary elections Sunday, but it was a lackluster victory that suggested the Russian leader's brand may be growing stale. United Russia, the pro-Kremlin behemoth an opposition leader once dubbed the "party of crooks and thieves," won less than 45% of the vote for 450 seats in the State Duma, Russia's lower house of parliament, according to preliminary results announced by the Central Election Committee on Sunday night. No other party came close. The nationalist LDPR party and the Communists competed for the second spot with 18% and 17% of the vote, respectively. Still, it was United Russia's lowest result in 15 years – and came in an election with turnout of less than 40% of eligible voters.

We establish general conditions under which Markov chains produced by the Hamiltonian Monte Carlo method will and will not be geometrically ergodic. We consider implementations with both position-independent and position-dependent integration times. In the former case we find that the conditions for geometric ergodicity are essentially a non-vanishing gradient of the log-density which asymptotically points towards the centre of the space and does not grow faster than linearly. In an idealised scenario in which the integration time is allowed to change in different regions of the space, we show that geometric ergodicity can be recovered for a much broader class of tail behaviours, leading to some guidelines for the choice of this free parameter in practice.

General predictive models do not provide a measure of confidence in predictions without Bayesian assumptions. A way to circumvent potential restrictions is to use conformal methods for constructing non-parametric confidence regions, that offer guarantees regarding validity. In this paper we provide a detailed description of a computationally efficient conformal procedure for Kernel Ridge Regression (KRR), and conduct a comparative numerical study to see how well conformal regions perform against the Bayesian confidence sets. The results suggest that conformalized KRR can yield predictive confidence regions with specified coverage rate, which is essential in constructing anomaly detection systems based on predictive models.

Principal component analysis (PCA) is a technique to find orthonormal vectors, which are a linear combination of the attributes of the data, that explain the variance structure of the data [12]. Since a few orthonormal vectors usually explain most of the variance, PCA is often used to reduce dimension of the data by keeping only a few of the orthonormal vectors. These orthonormal vectors are called principal components (PCs). For dimensionality reduction, we are given target dimension p, the number of PCs. To measure accuracy, given p principal components, first, the original data is projected into the lower dimension using the PCs. Next, the projected data in the lower dimension is lifted to the original dimension using the PCs. Observe that this procedure causes loss of some information if p is smaller than the dimension of the original attribute space. The reconstruction error is defined by the difference between the projected-and-lifted data and the original data. To select the best p PCs, the following two objective functions are usually used: [P1] minimization of the reconstruction error, [P2] maximization of the variance of the projected data.

As industry embraces AI, computers aren't the only ones that have to learn new tricks. Jen-Hsun Huang: 2015 was a big year. Artificial intelligence is moving into the commercial world. AI has been worked on for many years, largely in research. Various aspects of commercial use of AI, otherwise known as machine learning, is used for advertising and web searches and things like that. It wasn't until the last few years that AI could do things that people can't do. Several milestones were achieved in 2015 in particular that made it possible for us to use it in all kinds of areas. Yes, in an area of AI called "deep learning." The system basically learns by itself using a lot of data and computation.

MLSP 2016 - i.e. the IEEE International Workshop on Machine Learning for Signal Processing - was a great, well-organised workshop, held last week on Italy's Amalfi coast. Actually, there was substantial work involving Gaussian processes at MLSP. Is it a growth area? Well, if the use of GPs can be made more scalable (as in the above paper) then yes, it certainly should be. They are a very flexible and general tool, and nicely Bayesian too.

When it was first announced to a sceptical tech press months after a flop phone, the Echo was dismissed separately as a joke and a privacy nightmare. Now the latter may still prove to be the case (the Echo is always listening, and logs every sentence spoken to it), but a joke it is clearly not. In fact many analysts now believe that Amazon has one hand on the future that comes after the smartphone. Alexa is not the only, or even the first, voice-activated virtual assistant – Apple, Google and Microsoft have had their own for years – but it is the first that consumers have truly embraced. While taking out a smartphone in public and speaking to it – as one must with Apple's Siri or Google's Assistant – is awkward, and often slower than simply using a touchscreen, talking to a device in the comfort of one's own home is decidedly less uncomfortable.

Marketers will need to transform from campaigns to real-time, continuous interaction with customers via intelligent agents. So says Forrester Research VP and Principal Analyst Brian Hopkins, co-author (with Adam Silverman) of a new Forrester Research report, The Top Emerging Technologies to Watch: 2017 to 2021 ( 499 for individual purchase). It's about the top 15 developing technologies that will help businesses become more customer-obsessed over the next years. Forrester is obsessed with customer-obsession, which it says is essential to a modern brand and which is characterized by several key principles. According to the research firm, such a customer-focused company is led by insights from and about customers, responds quickly and is thoroughly connected everywhere. The report chose 15 technologies for their impact on companies employing these principles.