This paper provides a link between causal inference and machine learning techniques - specifically, Classification and Regression Trees (CART) - in observational studies where the receipt of the treatment is not randomized, but the assignment to the treatment can be assumed to be randomized (irregular assignment mechanism). The paper contributes to the growing applied machine learning literature on causal inference, by proposing a modified version of the Causal Tree (CT) algorithm to draw causal inference from an irregular assignment mechanism. The proposed method is developed by merging the CT approach with the instrumental variable framework to causal inference, hence the name Causal Tree with Instrumental Variable (CT-IV). As compared to CT, the main strength of CT-IV is that it can deal more efficiently with the heterogeneity of causal effects, as demonstrated by a series of numerical results obtained on synthetic data. Then, the proposed algorithm is used to evaluate a public policy implemented by the Tuscan Regional Administration (Italy), which aimed at easing the access to credit for small firms. In this context, CT-IV breaks fresh ground for target-based policies, identifying interesting heterogeneous causal effects.

This document aims to provide a review on learning with deep generative models (DGMs), which is an highly-active area in machine learning and more generally, artificial intelligence. This review is not meant to be a tutorial, but when necessary, we provide self-contained derivations for completeness. This review has two features. First, though there are different perspectives to classify DGMs, we choose to organize this review from the perspective of graphical modeling, because the learning methods for directed DGMs and undirected DGMs are fundamentally different. Second, we differentiate model definitions from model learning algorithms, since different learning algorithms can be applied to solve the learning problem on the same model, and an algorithm can be applied to learn different models. We thus separate model definition and model learning, with more emphasis on reviewing, differentiating and connecting different learning algorithms. We also discuss promising future research directions. This review is by no means comprehensive as the field is evolving rapidly. The authors apologize in advance for any missed papers and inaccuracies in descriptions. Corrections and comments are highly welcome.

We could initially model the problem as a machine learning or a deep learning problem. At this stage, we are concerned with the accuracy, choice and the efficiency of the model. Hence, the first quadrant is characterized by experimental analysis to prove value. We are also concerned with improving the existing KPIs. For example, if you are working with fraud detection or loan prediction – each of these applications has an existing KPI based on current techniques.

Thank you all for the huge response to this emerging course! We are delighted to have over 2300 students in over 102 different countries and for the overwhelmingly positive and thoughtful reviews. It's such a privilege to share this important topic with everyday people in a clear and understandable way. In this introductory course, the "Backyard Data Scientist" will guide you through wilderness of Machine Learning for Data Science. Accessible to everyone, this introductory course not only explains Machine Learning, but where it fits in the "techno sphere around us", why it's important now, and how it will dramatically change our world today and for days to come. We'll then explore the past and the future while touching on the importance, impacts and examples of Machine Learning for Data Science: To make sense of the Machine part of Machine Learning, we'll explore the Machine Learning process: Our final section of the course will prepare you to begin your future journey into Machine Learning for Data Science after the course is complete.

Artificial intelligence promises to make hiring an unbiased utopia. Employee referrals, a process that tends to leave underrepresented groups out, still make up a bulk of companies' hires. Recruiters and hiring managers also bring their own biases to the process, studies have found, often choosing people with the "right-sounding" names and educational background. Across the pipeline, companies lack racial and gender diversity, with the ranks of underrepresented people thinning at the highest levels of the corporate ladder. Fewer than 5 percent of chief executive officers at Fortune 500 companies are women, and that number will shrink further in October when Pepsi CEO Indra Nooyi steps down.

The concept of artificial intelligence and what it can do for education still remains elusive to many in the K-12 education space. A conversation I had recently with an assistant superintendent at the Colorado Association of School Executives convention underscored this idea. We began talking about artificial intelligence and district leader said, "You know, there's not a week that goes by that my superintendent isn't talking about AI!" But when I asked what the superintendent wanted to use AI for, the assistant superintendent just kind of looked at me with a raised eyebrow and shrugged. Artificial intelligence is in the water right now (some might say the Kool-Aid.) However, like many technical innovations from the past couple of decades, what it is and how it works is still a mystery to many people.

Modern machine learning algorithms have been adopted in a range of signal-processing applications spanning computer vision, natural language processing, and artificial intelligence. Many relevant problems involve subspace-structured features, orthogonality constrained or low-rank constrained objective functions, or subspace distances. These mathematical characteristics are expressed naturally using the Grassmann manifold. Unfortunately, this fact is not yet explored in many traditional learning algorithms. In the last few years, there have been growing interests in studying Grassmann manifold to tackle new learning problems. Such attempts have been reassured by substantial performance improvements in both classic learning and learning using deep neural networks. We term the former as shallow and the latter deep Grassmannian learning. The aim of this paper is to introduce the emerging area of Grassmannian learning by surveying common mathematical problems and primary solution approaches, and overviewing various applications. We hope to inspire practitioners in different fields to adopt the powerful tool of Grassmannian learning in their research.

We introduce the probably approximately correct (PAC) version of the problem of {Battling-bandits} with the Plackett-Luce (PL) model -- an online learning framework where in each trial, the learner chooses a subset of $k \le n$ arms from a pool of fixed set of $n$ arms, and subsequently observes a stochastic feedback indicating preference information over the items in the chosen subset; e.g., the most preferred item or ranking of the top $m$ most preferred items etc. The objective is to recover an `approximate-best' item of the underlying PL model with high probability. This framework is motivated by practical settings such as recommendation systems and information retrieval, where it is easier and more efficient to collect relative feedback for multiple arms at once. Our framework can be seen as a generalization of the well-studied PAC-{Dueling-Bandit} problem over set of $n$ arms. We propose two different feedback models: just the winner information (WI), and ranking of top-$m$ items (TR), for any $2\le m \le k$. We show that with just the winner information (WI), one cannot recover the `approximate-best' item with sample complexity lesser than $\Omega\bigg( \frac{n}{\epsilon^2} \ln \frac{1}{\delta}\bigg)$, which is independent of $k$, and same as the one required for standard dueling bandit setting ($k=2$). However with top-$m$ ranking (TR) feedback, our lower analysis proves an improved sample complexity guarantee of $\Omega\bigg( \frac{n}{m\epsilon^2} \ln \frac{1}{\delta}\bigg)$, which shows a relative improvement of $\frac{1}{m}$ factor compared to WI feedback, rightfully justifying the additional information gain due to the knowledge of ranking of topmost $m$ items. We also provide algorithms for each of the above feedback models, our theoretical analyses proves the {optimality} of their sample complexities which matches the derived lower bounds (upto logarithmic factors).

Andrew Yan-Tak Ng, regarded as one of the world's foremost experts on Artificial Intelligence (AI), firmly believes that despite the widespread mistrust of AI, it is good for governments, companies and individuals. Currently co-chairman and co-founder of the online learning platform Coursera and an adjunct professor at Stanford University's computer science department, Ng served as chief scientist and vice-president at Chinese tech company Baidu and was founding lead of the Google Brain team. In a phone interview from the Coursera headquarters in Mountain View, California, Ng spoke about the need for the Indian government to invest in education. He also shared his perspective on the potential of AI and the fears surrounding it. We would like you to propose one big idea to mark India's Independence Day.

In the last few years, machine learning applications have quietly entered every aspect of life: social media to speech recognition, radiology to retail, warfare to writing articles, coding to customer service, robotics to route optimization. During the 40 year information age, we told computers what to do. With advances in artificial intelligence, particularly machine learning, and faster processing chips we can feed computers giant data sets and they can (in narrow slivers) draw some inferences on their own. As we reported in Ask About AI, the rise of code that learns marks the beginning of a new era of augmented intelligence. It's a great opportunity for us to expand access to a great education and for young people to make a big contribution.