When it comes to developing drugs, the human brain is its own worst enemy. Though our pre-frontal cortex – the most complex structure known in nature – has devised countless remedies for the body that keeps it alive, the organ itself is almost impervious to treatment. The problem is that the filter system that stops toxins passing from the bloodstream into the brain also tends to block the very medicines designed to cure what ails it – from deadly tumours to Azlheimer's, Parkinson's and depression. Pharmaceutical companies have been hammering on this door for decades. Now Noor Shaker (pictured, above) wants to make the molecular equivalent of a skeleton key.

But a new system designed by Google DeepMind and British doctors goes a crucial step further: It can show users how it reached its conclusions. A study published Monday in Nature Medicine reports that the DeepMind system can identify dozens of diseases and point out the portions of optical coherence tomography scans that it relies upon to make its diagnoses. That's a crucial factor in validating the safety and efficacy of AI technologies being developed for use in diagnosing or recommending treatments for a broad range of diseases, from cancer to neurological and vision problems. The paper states that the system made the right referral recommendation in more than 94 percent of cases based on a review of historic patient scans at Moorfields Eye Hospital in London and performed as good as, or better than, top eye specialists who examined the same scans. Experts said that level of accuracy is impressive on such an open-ended query.

AI hype is at fever pitch. It's slated to fundamentally change everything about our world, from our economies to the way we get around cities. But how much of the hype is credible, and how much will AI change the nature of business in the near future? Will AI completely take over the realm of human expertise? Algorithms are changing the world.

"The human brain has 100 billion neurons, each neuron connected to 10 thousand other neurons. Sitting on your shoulders is the most complicated object in the known universe," Michio Kaku, Physicist and Futurist The human brain, which not just stores but also computes, is by far the most powerful and complex computers in the world that occupies just 1.3 litres of space and consumes about 20 watts of power. In comparison, the finest supercomputers in the world require gigawatts of power, massive real estate, infrastructure, and dedicated cooling systems while attempting to perform brain-like tasks. Understanding how the human brain functions and replicating it has been a lifelong quest for the scientific and research community. Enter neuromorphic computing, a concept developed by American scientist and researcher Carver Andress Mead in the late 1980s – which tries to emulate certain functions of the human brain in silicon.

Dell Technologies predicts the key IT trends for 2018. Driven by technologies such as Artificial Intelligence, Virtual and Augmented Reality and the Internet of Things, the deepening of cooperation between man and machine will drive positively the digitization of companies. In the next few years, companies will increasingly use the opportunity to let artificial intelligence (AI) think for themselves. In the AI systems, they set the parameters for classifying desired business outcomes, define the rules for their business activities, and set the framework for what constitutes an appropriate reward for their actions. Once these sets of rules are in place, the AI systems powered by data can show new business opportunities in near real time.

Despite the fact that nonlinear subspace learning techniques (e.g. manifold learning) have successfully applied to data representation, there is still room for improvement in explainability (explicit mapping), generalization (out-of-samples), and cost-effectiveness (linearization). To this end, a novel linearized subspace learning technique is developed in a joint and progressive way, called \textbf{j}oint and \textbf{p}rogressive \textbf{l}earning str\textbf{a}teg\textbf{y} (J-Play), with its application to multi-label classification. The J-Play learns high-level and semantically meaningful feature representation from high-dimensional data by 1) jointly performing multiple subspace learning and classification to find a latent subspace where samples are expected to be better classified; 2) progressively learning multi-coupled projections to linearly approach the optimal mapping bridging the original space with the most discriminative subspace; 3) locally embedding manifold structure in each learnable latent subspace. Extensive experiments are performed to demonstrate the superiority and effectiveness of the proposed method in comparison with previous state-of-the-art methods.

Non-Gaussian component analysis (NGCA) is a problem in multidimensional data analysis. Since its formulation in 2006, NGCA has attracted considerable attention in statistics and machine learning. In this problem, we have a random variable $X$ in $n$-dimensional Euclidean space. There is an unknown subspace $U$ of the $n$-dimensional Euclidean space such that the orthogonal projection of $X$ onto $U$ is standard multidimensional Gaussian and the orthogonal projection of $X$ onto $V$, the orthogonal complement of $U$, is non-Gaussian, in the sense that all its one-dimensional marginals are different from the Gaussian in a certain metric defined in terms of moments. The NGCA problem is to approximate the non-Gaussian subspace $V$ given samples of $X$. Vectors in $V$ corresponds to "interesting" directions, whereas vectors in $U$ correspond to the directions where data is very noisy. The most interesting applications of the NGCA model is for the case when the magnitude of the noise is comparable to that of the true signal, a setting in which traditional noise reduction techniques such as PCA don't apply directly. NGCA is also related to dimensionality reduction and to other data analysis problems such as ICA. NGCA-like problems have been studied in statistics for a long time using techniques such as projection pursuit. We give an algorithm that takes polynomial time in the dimension $n$ and has an inverse polynomial dependence on the error parameter measuring the angle distance between the non-Gaussian subspace and the subspace output by the algorithm. Our algorithm is based on relative entropy as the contrast function and fits under the projection pursuit framework. The techniques we develop for analyzing our algorithm maybe of use for other related problems.

From self-driving vehicles and back-flipping robots to virtual assistants who book our next appointment at the hair salon or at that restaurant for dinner - machine learning systems are becoming increasingly ubiquitous. The main reason for this is that these methods boast remarkable predictive capabilities. However, most of these models remain black boxes, meaning that it is very challenging for humans to follow and understand their intricate inner workings. Consequently, interpretability has suffered under this ever-increasing complexity of machine learning models. Especially with regards to new regulations, such as the General Data Protection Regulation (GDPR), the necessity for plausibility and verifiability of predictions made by these black boxes is indispensable. Driven by the needs of industry and practice, the research community has recognised this interpretability problem and focussed on developing a growing number of so-called explanation methods over the past few years. These methods explain individual predictions made by black box machine learning models and help to recover some of the lost interpretability. With the proliferation of these explanation methods, it is, however, often unclear, which explanation method offers a higher explanation quality, or is generally better-suited for the situation at hand. In this thesis, we thus propose an axiomatic framework, which allows comparing the quality of different explanation methods amongst each other. Through experimental validation, we find that the developed framework is useful to assess the explanation quality of different explanation methods and reach conclusions that are consistent with independent research.

The presence of Long Distance Dependencies (LDDs) in sequential data poses significant challenges for computational models. Various recurrent neural architectures have been designed to mitigate this issue. In order to test these state-of-the-art architectures, there is growing need for rich benchmarking datasets. However, one of the drawbacks of existing datasets is the lack of experimental control with regards to the presence and/or degree of LDDs. This lack of control limits the analysis of model performance in relation to the specific challenge posed by LDDs. One way to address this is to use synthetic data having the properties of subregular languages. The degree of LDDs within the generated data can be controlled through the k parameter, length of the generated strings, and by choosing appropriate forbidden strings. In this paper, we explore the capacity of different RNN extensions to model LDDs, by evaluating these models on a sequence of SPk synthesized datasets, where each subsequent dataset exhibits a longer degree of LDD. Even though SPk are simple languages, the presence of LDDs does have significant impact on the performance of recurrent neural architectures, thus making them prime candidate in benchmarking tasks.

Modeling spillover effects from observational data is an important problem in economics, business, and other fields of research. % It helps us infer the causality between two seemingly unrelated set of events. For example, if consumer spending in the United States declines, it has spillover effects on economies that depend on the U.S. as their largest export market. In this paper, we aim to infer the causation that results in spillover effects between pairs of entities (or units), we call this effect as \textit{paired spillover}. To achieve this, we leverage the recent developments in variational inference and deep learning techniques to propose a generative model called Linked Causal Variational Autoencoder (LCVA). Similar to variational autoencoders (VAE), LCVA incorporates an encoder neural network to learn the latent attributes and a decoder network to reconstruct the inputs. However, unlike VAE, LCVA treats the \textit{latent attributes as confounders that are assumed to affect both the treatment and the outcome of units}. Specifically, given a pair of units $u$ and $\bar{u}$, their individual treatment and outcomes, the encoder network of LCVA samples the confounders by conditioning on the observed covariates of $u$, the treatments of both $u$ and $\bar{u}$ and the outcome of $u$. Once inferred, the latent attributes (or confounders) of $u$ captures the spillover effect of $\bar{u}$ on $u$. Using a network of users from job training dataset (LaLonde (1986)) and co-purchase dataset from Amazon e-commerce domain, we show that LCVA is significantly more robust than existing methods in capturing spillover effects.