Artificial Intelligence (AI) is nothing new. The field of AI research was founded in 1956. To date, this field has been always covered with huge expectations. Surely, we are in such a hype phase, but Google CEO Sundar Pichai is also right, when he says AI is bigger than the invention of fire and electricity. As one of the largest industrial nations, Germany addresses this big promise and published a AI strategy this month (Nov, 2018).

I want to take a moment to tell the origin story of regression analysis, which will explain why it has that name. I believe that of all the common machine learning techniques (K-means, kNN, PCA), "regression analysis" has the most opaque name. OLS regression was first invented to analyze exceptional genetic traits and their heritability. These early studies seemed to show the offspring of exceptional individuals "regressed to the mean". The inventor was Sir Francis Galton (half-cousin of Charles Darwin²), who had previously invented the standard deviation and first observed the "wisdom of the crowds" in certain estimation tasks. I am trying to predict daily demand for short-term bike rentals made in 2012, and I have data from 2011 to build the model.

Don't simply "all in" if there's a bot at your Texas hold'em poker table, because Facebook and Carnegie Mellon University's new Pluribus AI system just beat five human pros at the same time -- including a couple of World Series of Poker Champs. AI models had already bettered human poker pros one-on-one, but Pluribus's success in a six-player game signals a huge leap in ability. Texas hold'em is one of the most popular poker variants that involves game theory, gambling, and strategy. To win the game, each play must assemble the best five cards from any combination of two "hole cards" dealt face down to each player and five community cards dealt face up. Players can choose to check, bet, call, raise, and fold. Researchers regard poker as a meaningful and complex experimental field where they can explore how AI interacts with gaming theory and imperfect information.

AI will be a major growth driver for Asia in the coming decade. The company priorities for AI are to enhance customer satisfaction, speed up decision-making, and reduce inefficiencies. The loss of some roles to automation, and the restructuring of others to take advantage of technology-created capacity, are likely. Yet reducing headcount is not a top priority in and of itself. Just one-third of survey respondents listed the need to reduce labor costs as a top-three driver for AI.

In Q2 2018, Chinese investors sank $2.87b into AI startups; in Q2 2019, it was $140.7m. It's part of a massive slowdown in China's AI industry, which kicked off with massive political/economic fanfare from the Chinese state, which promised that the sector would be worth $150b by 2030, a boast that touched off anxiety about a global AI arms race. Two years later, valuations for the companies that bet biggest on AI are plummeting. Baidu has sunk to a valuation of 10% of the worth of rivals Alibaba and Tencent, though they were all level as recently as 2017; the company's top AI scientists have quit, and the company just booked its first losses since 2005. Many of the early promising AI demos have fizzled or turned out to be smoke-and-mirrors: much-vaunted demonstrations of health-tech companies like Ping An to diagnose diseases early and head them off before they could spread represent mere incremental improvements over techniques that were documented and demonstrated in the 1970s.

Three teams have developed artificial muscles that can lift 1000 times their own weight. They hope the new fibres could be used in prosthetic limbs, robots, exoskeletons, and even in clothing. All three teams have developed their muscles according to a similar principle: that a coiled-up substance can stretch like a muscle. The idea was developed by Ray Baughman and his colleagues at the University of Texas, who found that twisting up even a simple material like sewing thread or fishing line can create a muscle-like structure that, for its size, can lift weights 100 times heavier than human muscle can manage. Now, Baughman's team have developed stronger fibres, using similarly inexpensive materials.

Artificial intelligence has finally cracked the biggest challenge in poker: beating top professionals in six-player no-limit Texas Hold'Em, the most popular variant of the game. Over 20,000 hands of online poker, the AI beat fifteen of the world's top poker players, each of whom has won more than $1 million USD playing the game professionally. The AI, called Pluribus, was tested in 10,000 games against five human players, as well as in 10,000 rounds where five copies of Pluribus played against one professional – and did better than the pros in both. Pluribus was developed by Noam Brown of Facebook AI Research and Tuomas Sandholm at Carnegie Mellon University in the US. It is an improvement on their previous poker-playing AI, called Libratus, which in 2017 outplayed professionals at Heads-Up Texas Hold'Em, a variant of the game that pits two players head to head.

An artificial intelligence called Pluribus has emerged victorious from a marathon 12-day poker session during which it played five human professionals at a time. Over 10,000 hands of no-limit Texas hold'em, the most popular form of the game, Pluribus won a virtual $48,000 (£38,000), beating five elite players who were selected each day from a pool who agreed to take on the program. All of the pros had previously won more than $1m playing the game. What counts as a beating for humanity ranks as a milestone for AI. No computer program has ever achieved superhuman performance against multiple poker players. A forerunner of Pluribus named Libratus made its name two years ago by trouncing top human players, but that program only played one-on-one.

Previous literature on unsupervised learning focused on designing structural priors and optimization functions with the aim of learning meaningful features, but without considering the description length of the representations. Here we present Sparsely Activated Networks (SANs), which decompose their input as a sum of sparsely reoccurring patterns of varying amplitude, and combined with a newly proposed metric $\varphi$ they learn representations with minimal description lengths. SANs consist of kernels with shared weights that during encoding are convolved with the input and then passed through a ReLU and a sparse activation function. During decoding, the same weights are convolved with the sparse activation map and the individual reconstructions from each weight are summed to reconstruct the input. We also propose a metric $\varphi$ for model selection that favors models which combine high compression ratio and low reconstruction error and we justify its definition by exploring the hyperparameter space of SANs. We compare four sparse activation functions (Identity, Max-Activations, Max-Pool indices, Peaks) on a variety of datasets and show that SANs learn interpretable kernels that combined with $\varphi$, they minimize the description length of the representations.

The Gaussian process (GP) is a nonparametric prior distribution over functions indexed by time, space, or other high-dimensional index set. The GP is a flexible model yet its limitation is given by its very nature: it can only model Gaussian marginal distributions. To model non-Gaussian data, a GP can be warped by a nonlinear transformation (or warping) as performed by warped GPs (WGPs) and more computationally-demanding alternatives such as Bayesian WGPs and deep GPs. However, the WGP requires a numerical approximation of the inverse warping for prediction, which increases the computational complexity in practice. To sidestep this issue, we construct a novel class of warpings consisting of compositions of multiple elementary functions, for which the inverse is known explicitly. We then propose the compositionally-warped GP (CWGP), a non-Gaussian generative model whose expressiveness follows from its deep compositional architecture, and its computational efficiency is guaranteed by the analytical inverse warping. Experimental validation using synthetic and real-world datasets confirms that the proposed CWGP is robust to the choice of warpings and provides more accurate point predictions, better trained models and shorter computation times than WGP.