Looking at the latest Google and Apple conventions, it is clear to all: If in the past years the main buzzwords in the information technology field were IoT and Big Data, the catch-all word of this year is without a doubt Machine Learning. What does this word mean exactly? Are we talking about Artificial Intelligence? Is somebody trying to build a Skynet to ruin the world? Will machines steal my job in the future?

We study the fundamental problem of high-dimensional mean estimation in a robust model where a constant fraction of the samples are adversarially corrupted. Recent work gave the first polynomial time algorithms for this problem with dimension-independent error guarantees for several families of structured distributions. In this work, we give the first nearly-linear time algorithms for high-dimensional robust mean estimation. Specifically, we focus on distributions with (i) known covariance and sub-gaussian tails, and (ii) unknown bounded covariance. Given $N$ samples on $\mathbb{R}^d$, an $\epsilon$-fraction of which may be arbitrarily corrupted, our algorithms run in time $\tilde{O}(Nd) / \mathrm{poly}(\epsilon)$ and approximate the true mean within the information-theoretically optimal error, up to constant factors. Previous robust algorithms with comparable error guarantees have running times $\tilde{\Omega}(N d^2)$, for $\epsilon = \Omega(1)$. Our algorithms rely on a natural family of SDPs parameterized by our current guess $\nu$ for the unknown mean $\mu^\star$. We give a win-win analysis establishing the following: either a near-optimal solution to the primal SDP yields a good candidate for $\mu^\star$ -- independent of our current guess $\nu$ -- or the dual SDP yields a new guess $\nu'$ whose distance from $\mu^\star$ is smaller by a constant factor. We exploit the special structure of the corresponding SDPs to show that they are approximately solvable in nearly-linear time. Our approach is quite general, and we believe it can also be applied to obtain nearly-linear time algorithms for other high-dimensional robust learning problems.

Digital Health is being revolutionised by Mathematical Models, Machine Learning (ML) and Artificial Intelligence (AI). This makes it quicker and easier for therapists and possible for everyday people to interpret complex medical data. AI and ML can be used to help identify, predict and prevent disease before it happens. To do this, the Digital Health Industry is focused on developing and using Software as a Technology and as a Service. We would like to introduce you to two startups who are doing just that.

In real-world applications such as emotion recognition from recorded brain activity, data are captured from electrodes over time. These signals constitute a multidimensional time series. In this paper, Echo State Network (ESN), a recurrent neural network with a great success in time series prediction and classification, is optimized with different neural plasticity rules for classification of emotions based on electroencephalogram (EEG) time series. Actually, the neural plasticity rules are a kind of unsupervised learning adapted for the reservoir, i.e. the hidden layer of ESN. More specifically, an investigation of Oja's rule, BCM rule and gaussian intrinsic plasticity rule was carried out in the context of EEG-based emotion recognition. The study, also, includes a comparison of the offline and online training of the ESN. When testing on the well-known affective benchmark "DEAP dataset" which contains EEG signals from 32 subjects, we find that pretraining ESN with gaussian intrinsic plasticity enhanced the classification accuracy and outperformed the results achieved with an ESN pretrained with synaptic plasticity. Four classification problems were conducted in which the system complexity is increased and the discrimination is more challenging, i.e. inter-subject emotion discrimination. Our proposed method achieves higher performance over the state of the art methods.

We study the effect of impairment on stochastic multi-armed bandits and develop new ways to mitigate it. Impairment effect is the phenomena where an agent only accrues reward for an action if they have played it at least a few times in the recent past. It is practically motivated by repetition and recency effects in domains such as advertising (here consumer behavior may require repeat actions by advertisers) and vocational training (here actions are complex skills that can only be mastered with repetition to get a payoff). Impairment can be naturally modelled as a temporal constraint on the strategy space, and we provide two novel algorithms that achieve sublinear regret, each working with different assumptions on the impairment effect. We introduce a new notion called bucketing in our algorithm design, and show how it can effectively address impairment as well as a broader class of temporal constraints. Our regret bounds explicitly capture the cost of impairment and show that it scales (sub-)linearly with the degree of impairment. Our work complements recent work on modeling delays and corruptions, and we provide experimental evidence supporting our claims.

The moment a hyped-up new technology garners mainstream attention, many businesses will scramble to incorporate it into their enterprise. The majority of these trends will splutter and die out by Q4. Artificial intelligence (AI) is unlikely to be one of them. AI is a transformative series of tools that can accelerate productivity, drive insight, and open up unexplored revenue streams. It's poised to revolutionize the way we do business and everyone in a leadership role should be thinking about it.

Business person Brian Beaumont has overcome challenges brought on by dyslexia. Brian Beaumont was a below average student prior to entering graduate school in the early 1980s. So Beaumont, now 60, asked his professors if he could tape their lectures to make better use of his 60- to 90-minute commute time in and around Los Angeles. "I did not realize at the time I was making an accommodation for my dyslexia," Beaumont says. "I had problems listening and taking notes at the same time. Now, I could sit back and just listen to the lecture. I could focus on the main points the professor was making."

Stanford graduate student Kaidi Cao will join fellow AI researchers Jing Liao, of City University of Hong Kong, and Lu Yuan of Microsoft at SIGGRAPH Asia in Tokyo this December to present their incredible caricature-drawing neural network. That's not bad, considering Cao was only an intern at the Visual Computing Group at the Microsoft Research Lab in Beijing when he worked on the project. The AI, actually a pair of generative adversarial networks (GAN), is called CariGANs. The first of its neural networks, CariGeoGAN, determines the geometry of a face in a photograph and maps it to a caricature model. CariStyGAN, the other half of CariGANs, does the "style transfer," or applies the artistic look to the geometry map.

It aims to provide students with an understanding of the role computation can play in solving problems and to help students, regardless of their major, feel justifiably confident of their ability to write small programs that allow them to accomplish useful goals. The class uses the Python 3.5 programming language.

Safe, controllable and credible artificial intelligence has been the goal which the humanity has been pursuing. In the field of machine learning, in order to achieve this goal, it is necessary for learning algorithm to really interact with the humanity; It is necessary for the learning algorithm to have the ability to correct errors, so as to avoid a prediction model with serious errors caused by unnecessary deviation in training data; It needs to be able to check its own learning process or decision-making process based on unsuccessful prediction results, especially for complex learning tasks; It is necessary to establish a learning algorithm for capturing and learning causal relationships in the world around us, so that the prediction model could predict what will happen under certain conditions, even if these conditions are significantly different from those of the past; It needs the learning algorithm which can really take full control of generalization performance of the prediction model. As big data accelerates transformation of scientific research pattern, scientific research is translating from a hypothetical drive mode to a data-driven one, which needs learning algorithm to discover new natural phenomena and laws through big data mining, statistic and analysis. However, recently, all of this is out of reach. The reason is that the prediction model and its training process are not yet understood by human beings, and are not covered by the knowledge base we currently have.