We consider a repeated newsvendor problem where the inventory manager has no prior information about the demand, and can access only censored/sales data. In analogy to multi-armed bandit problems, the manager needs to simultaneously "explore" and "exploit" with her inventory decisions, in order to minimize the cumulative cost. We make no probabilistic assumptions---importantly, independence or time stationarity---regarding the mechanism that creates the demand sequence. Our goal is to shed light on the hardness of the problem, and to develop policies that perform well with respect to the regret criterion, that is, the difference between the cumulative cost of a policy and that of the best fixed action/static inventory decision in hindsight, uniformly over all feasible demand sequences. We show that a simple randomized policy, termed the Exponentially Weighted Forecaster, combined with a carefully designed cost estimator, achieves optimal scaling of the expected regret (up to logarithmic factors) with respect to all three key primitives: the number of time periods, the number of inventory decisions available, and the demand support. Through this result, we derive an important insight: the benefit from "information stalking" as well as the cost of censoring are both negligible in this dynamic learning problem, at least with respect to the regret criterion. Furthermore, we modify the proposed policy in order to perform well in terms of the tracking regret, that is, using as benchmark the best sequence of inventory decisions that switches a limited number of times. Numerical experiments suggest that the proposed approach outperforms existing ones (that are tailored to, or facilitated by, time stationarity) on nonstationary demand models. Finally, we extend the proposed approach and its analysis to a "combinatorial" version of the repeated newsvendor problem.

Classifiers and rating scores are prone to implicitly codifying biases, which may be present in the training data, against protected classes (i.e., age, gender, or race). So it is important to understand how to design classifiers and scores that prevent discrimination in predictions. This paper develops computationally tractable algorithms for designing accurate but fair support vector machines (SVM's). Our approach imposes a constraint on the covariance matrices conditioned on each protected class, which leads to a nonconvex quadratic constraint in the SVM formulation. We develop iterative algorithms to compute fair linear and kernel SVM's, which solve a sequence of relaxations constructed using a spectral decomposition of the nonconvex constraint. Its effectiveness in achieving high prediction accuracy while ensuring fairness is shown through numerical experiments on several data sets.

New social and economic activities massively exploit big data and machine learning algorithms to do inference on people's lives. Applications include automatic curricula evaluation, wage determination, and risk assessment for credits and loans. Recently, many governments and institutions have raised concerns about the lack of fairness, equity and ethics in machine learning to treat these problems. It has been shown that not including sensitive features that bias fairness, such as gender or race, is not enough to mitigate the discrimination when other related features are included. Instead, including fairness in the objective function has been shown to be more efficient. We present novel fair regression and dimensionality reduction methods built on a previously proposed fair classification framework. Both methods rely on using the Hilbert Schmidt independence criterion as the fairness term. Unlike previous approaches, this allows us to simplify the problem and to use multiple sensitive variables simultaneously. Replacing the linear formulation by kernel functions allows the methods to deal with nonlinear problems. For both linear and nonlinear formulations the solution reduces to solving simple matrix inversions or generalized eigenvalue problems. This simplifies the evaluation of the solutions for different trade-off values between the predictive error and fairness terms. We illustrate the usefulness of the proposed methods in toy examples, and evaluate their performance on real world datasets to predict income using gender and/or race discrimination as sensitive variables, and contraceptive method prediction under demographic and socio-economic sensitive descriptors.

Impressive advances in artificial intelligence technology tailored for legal work have led some lawyers to worry that their profession may be Silicon Valley's next victim. But recent research and even the people working on the software meant to automate legal work say the adoption of A.I. in law firms will be a slow, task-by-task process. In other words, like it or not, a robot is not about to replace your lawyer. "There is this popular view that if you can automate one piece of the work, the rest of the job is toast," said Frank Levy, a labor economist at the Massachusetts Institute of Technology. An artificial intelligence technique called natural language processing has proved useful in scanning and predicting what documents will be relevant to a case, for example. Yet other lawyers' tasks, like advising clients, writing legal briefs, negotiating and appearing in court, seem beyond the reach of computerization, for a while.

For three years, I've assisted the American Accounting Association in the development of their Accounting IS Big Data show. It's been fascinating to watch accounting professionals and academics change their perspective as the evolution of accounting as new technologies come into view. It's a process of that follows the classic denial, resist and acceptance pattern we see in technology adoption. Here's a recap of what's happened. Attendees at the inaugural show, were in for a shock.

Artificial intelligence technologies might find a considerable hurdle in privacy rights, obligations and sanctions provided by the EU General Data Protection Regulation. As part of the series of blog posts on the major changes introduced by the EU Data Protection Regulation, here is an article on how it might affect artificial intelligence. Artificial intelligence (AI) is intelligence exhibited by machines. In computer science, an ideal "intelligent" machine is a flexible rational agent that perceives its environment and takes actions that maximize its chance of success at some goal. Colloquially, the term "artificial intelligence" is applied when a machine mimics "cognitive" functions that humans associate with other human minds, such as "learning" and "problem solving".

The mass-scale deployment of robots has already ushered in a whole new world of work. It's a CEO's capitalist paradise, where the workforce doesn't call in sick or take vacations, can't file lawsuits, doesn't organize unions, and is cheap. As a result, robots are rapidly climbing the pay ladder into white-collar and professional positions that millions of college-educated, middle-class employees have wrongly considered safe....

As we all know, there are pros and cons associated with every technology -- and AI (artificial intelligence) is no exception to this rule. The most popular are dating bots, where a computer program (chatbot) that uses artificial intelligence strikes up conversations with dating site users, enabling the scammer to "talk" with multiple potential victims at once. Legal challenges related to AI's application in the financial industry could be related to the consequences of erroneous algorithms and data governance. Rather, organizations should focus on how they can responsibly reduce the ill effects of this technology by minimizing the challenges and leveraging the benefits and by creating a clear technology adoption roadmap that understands its core capabilities.

As we all know, there are pros and cons associated with every technology -- and AI (artificial intelligence) is no exception to this rule. However, it becomes very important for us as technology consumers and producers to identify those challenges and minimize the associated risks and at the same time make sure that we take full advantage of this technology. In this article, I will break down the potential challenges that AI developers need to address to make sure that AI is accepted as a friend and not as a foe, given the kind of media attention it is getting -- especially around the fact that it is going to take away our jobs. The widespread use of AI raised a number of issues that have yet to be addressed. As we know, AI bots are becoming better at mimicking human conversations. For example, in 2015, a bot named Eugene Goostman won the Turing Challenge for the first time.

ARTICLE 19 submitted evidence to the United Kingdom's House of Lords Select Committee on Artificial Intelligence on 6 September 2017. The submission stresses the need to critically evaluate the impact of Artificial Intelligence (AI) and automated decision making systems (AS) on human rights. It also calls for deeper understanding of various ways in which these technologies embed values and bias, thereby strengthening or sometimes hindering the exercise of these rights, particularly freedom of expression. The overarching recommendation is for the development and use of AI to be subject to the minimum requirement of respecting, promoting, and protecting international human rights standards. Since 2014, ARTICLE 19 has pioneered efforts in technical communities to bridge existing knowledge gaps on human rights and their relevance in internet infrastructure.