The year 2017 saw artificial intelligence bringing the stuff of science fiction closer to reality by not only gaining foothold in all spheres of life, but also getting the better of humans in many fields. From acquiring citizenship to outsmarting humans at complex games, from composing music to writing novels, from assisting doctors to helping fight judicial cases, artificial intelligence (AI) made its presence felt throughout the year. Artificial intelligence is a term used to describe systems or machines that mimic the cognitive functions of human minds, such as learning and problem solving. Although by no means a new concept, the technology made headlines throughout the year. Perhaps, among the most talked about AI machines this year was Sophia, a humanoid robot designed by a company in Hong Kong, that was granted citizenship in Saudi Arabia - a country where women were not allowed to drive until recently.

Predictive modelling and supervised learning are central to modern data science. With predictions from an ever-expanding number of supervised black-box strategies - e.g., kernel methods, random forests, deep learning aka neural networks - being employed as a basis for decision making processes, it is crucial to understand the statistical uncertainty associated with these predictions. As a general means to approach the issue, we present an overarching framework for black-box prediction strategies that not only predict the target but also their own predictions' uncertainty. Moreover, the framework allows for fair assessment and comparison of disparate prediction strategies. For this, we formally consider strategies capable of predicting full distributions from feature variables, so-called probabilistic supervised learning strategies. Our work draws from prior work including Bayesian statistics, information theory, and modern supervised machine learning, and in a novel synthesis leads to (a) new theoretical insights such as a probabilistic bias-variance decomposition and an entropic formulation of prediction, as well as to (b) new algorithms and meta-algorithms, such as composite prediction strategies, probabilistic boosting and bagging, and a probabilistic predictive independence test. Our black-box formulation also leads (c) to a new modular interface view on probabilistic supervised learning and a modelling workflow API design, which we have implemented in the newly released skpro machine learning toolbox, extending the familiar modelling interface and meta-modelling functionality of sklearn. The skpro package provides interfaces for construction, composition, and tuning of probabilistic supervised learning strategies, together with orchestration features for validation and comparison of any such strategy - be it frequentist, Bayesian, or other.

Although deep learning has historical roots going back decades, neither the term "deep learning" nor the approach was popular just over five years ago, when the field was reignited by papers such as Krizhevsky, Sutskever and Hinton's now classic 2012 (Krizhevsky, Sutskever, & Hinton, 2012)deep net model of Imagenet. What has the field discovered in the five subsequent years? Against a background of considerable progress in areas such as speech recognition, image recognition, and game playing, and considerable enthusiasm in the popular press, I present ten concerns for deep learning, and suggest that deep learning must be supplemented by other techniques if we are to reach artificial general intelligence.

Transfer learning leverages the knowledge in one domain, the source domain, to improve learning efficiency in another domain, the target domain. Existing transfer learning research is relatively well-progressed, but only in situations where the feature spaces of the domains are homogeneous and the target domain contains at least a few labeled instances. However, transfer learning has not been well-studied in heterogeneous settings with an unlabeled target domain. To contribute to the research in this emerging field, this paper presents: (1) an unsupervised knowledge transfer theorem that prevents negative transfer; and (2) a principal angle-based metric to measure the distance between two pairs of domains. The metric shows the extent to which homogeneous representations have preserved the information in original source and target domains. The unsupervised knowledge transfer theorem sets out the transfer conditions necessary to prevent negative transfer. Linear monotonic maps meet the transfer conditions of the theorem and, hence, are used to construct homogeneous representations of the heterogeneous domains, which in principle prevents negative transfer. The metric and the theorem have been implemented in an innovative transfer model, called a Grassmann-LMM-geodesic flow kernel (GLG), that is specifically designed for knowledge transfer across heterogeneous domains. The GLG model learns homogeneous representations of heterogeneous domains by minimizing the proposed metric. Knowledge is transferred through these learned representations via a geodesic flow kernel. Notably, the theorem presented in this paper provides the sufficient transfer conditions needed to guarantee that knowledge is transferred from a source domain to an unlabeled target domain with correctness.

This paper describes a suite of algorithms for constructing low-rank approximations of an input matrix from a random linear image of the matrix, called a sketch. These methods can preserve structural properties of the input matrix, such as positive-semidefiniteness, and they can produce approximations with a user-specified rank. The algorithms are simple, accurate, numerically stable, and provably correct. Moreover, each method is accompanied by an informative error bound that allows users to select parameters a priori to achieve a given approximation quality. These claims are supported by numerical experiments with real and synthetic data.

Keywords: EEG, electroencephalogram, seizure detection, machine learning The electroencephalogram (EEG), which has been in clinical use for over 70 years, is still an essential tool for diagnosis of neural functioning (Kennett, 2012). Well-known applications of EEGs include identification of epilepsy and epileptic seizures, anoxic and hypoxic damage to the brain, and identification of neural disorders such as hemorrhagic stroke, ischemia and toxic metabolic encephalopathy (Drury, 1988). More recently there has been interest in diagnosing Alzheimer's (Tsolaki et al., 2014), head trauma (Rapp et al., 2015) and sleep disorders (Younes, 2017). Many of these clinical applications now involve the collection of large amounts of data (e.g., 72-hour continuous EEG recordings), which makes manual interpretation challenging. Similarly, the increased use of EEGs in critical care has created a significant demand for high-performance automatic interpretation software (e.g., real-time seizure detection).

It won't be long before the computers put us all out of jobs. For the medical field, it looks like some of us will be out of a job before others. An exciting study from the Netherlands utilized computerized analysis of pathologic slides to look for metastatic breast cancer in lymph nodes in specimens from women with known breast cancer. Finding metastatic cancer in a lymph node is an important indicator for recurrence of disease if untreated, a prognostic sign, and a reason to undergo more extensive treatment. Obviously, doing whatever we can to accurately and thoroughly assess such tissue is extremely important.

We are confronted with reports and research on an increasing number of impressively accurate IT tools that interpret images or ECG data, that advise on when to consult a doctor or that suggest proper therapies. The year of 2018 will be when these tools start hitting real-world care on a broader scale. The FDA has just certified the first Apple Watch accessory as a medical device: an ECG-meter that replaces the normal Apple watch wristband and that can be used in combination with activity tracking in a machine-learning environment to detect atrial fibrillation. Decision support tools with or without machine learning are coming of age, and this will force regulators to make up their minds on how to address them properly. The recent European Medical Device Regulation will make it tougher for healthcare IT tools that need medicinal product certification to make ends meet.

But such talk has little currency in Sweden or its Scandinavian neighbors, where unions are powerful, government support is abundant, and trust between employers and employees runs deep. Here, robots are just another way to make companies more efficient. As employers prosper, workers have consistently gained a proportionate slice of the spoils -- a stark contrast to the United States and Britain, where wages have stagnated even while corporate profits have soared. "In Sweden, if you ask a union leader, 'Are you afraid of new technology?' they will answer, 'No, I'm afraid of old technology,'" says the Swedish minister for employment and integration, Ylva Johansson. "The jobs disappear, and then we train people for new jobs. But we will protect workers."

This year marks the 200th anniversary of the publication of Mary Shelley's classic novel Frankenstein - first printed on 1 January 1818. Shelley came up with the idea at the age of 18 after being challenged by romantic poet Lord Byron, while in Switzerland, to construct a ghost story. The results were to have a monumental impact. This was the kernel from which the story of Frankenstein would emerge. The novel - originally published without Shelley's name - received mixed reviews, but came into prominence after being picked up and re-versioned by theatre companies a few years later.