Data science has been dubbed the coolest job of the 21st century -- beating out a tough field including game developer and professional sportsperson -- but is data science even a profession yet? Clearly, data scientist as a job is one of the new kids on the block, which is a cause of some of the coolness -- but may not be so great after five years or twenty years. How far has data science developed as a profession? When can we properly call ourselves a profession, and expect the kind of recognition ( way beyond coolness) that surgeons and lawyers expect? Within data science there is the beginnings of an understanding of which skills should be considered core, but not it stops a long way short of being shared to the extent found in other professions.

In 2016 The Washington Post unleashed a new reporter on the world, an artificial intelligence (AI) system called Heliograf. In its first year, it churned out 300 short reports on the Rio Olympics, followed by 500 brief articles about the presidential election, which clocked up pretty good engagement online. Meanwhile, pharmaceutical companies are increasingly turning to AI to drastically speed up the process of discovering new drugs, analysing huge quantities of data to come up with new molecules that could potentially have a therapeutic effect. However, according to most legal and technology experts, this scenario is a long way off. "From my perspective, at present AI is little more than a tool that can be wielded by the creator of a creative work or inventor of a new technical innovation in the same way a paintbrush is wielded by an artist or a CAD [computer-aided design] tool by an inventor," says Jeremy Smith, chartered patent attorney and partner at IP law firm Mathys & Squire.

Inside a sunny conference room on the Microsoft campus in Redmond, Wash., a small team of employees is describing how technology can save the world. Microsoft's Digital Diplomacy unit consists of two dozen policy experts who work on everything from the ethical use of artificial intelligence to protecting the 2020 presidential election from foreign cyberinterference. Brad Smith, Microsoft's president, sits in the middle of the table, sipping coffee from a mug bearing the name of his hometown, Appleton, Wis. The group updates Smith on a tech-industry initiative co-founded by Microsoft to combat terrorist messaging on the Internet. Smith pushes for more ideas. "We need something that will create a new mold," he says.

The widespread use of machine learning models, especially within the context of decision-making systems impacting individuals, raises many ethical issues with respect to fairness and interpretability of these models. While the research in these domains is booming, very few works have addressed these two issues simultaneously. To solve this shortcoming, we propose FairCORELS, a supervised learning algorithm whose objective is to learn at the same time fair and interpretable models. FairCORELS is a multi-objective variant of CORELS, a branch-and-bound algorithm, designed to compute accurate and interpretable rule lists. By jointly addressing fairness and interpretability, FairCORELS can achieve better fairness/accuracy tradeoffs compared to existing methods, as demonstrated by the empirical evaluation performed on real datasets. Our paper also contains additional contributions regarding the search strategies for optimizing the multi-objective function integrating both fairness, accuracy and interpretability.

This paper develops a communication-efficient algorithm to solve the stochastic optimization problem defined over a distributed network, aiming at reducing the burdensome communication in applications such as distributed machine learning. Different from the existing works based on quantization and sparsification, we introduce a communication-censoring technique to reduce the transmissions of variables, which leads to our communication-Censored distributed Stochastic Gradient Descent (CSGD) algorithm. Specifically, in CSGD, the latest mini-batch stochastic gradient at a worker will be transmitted to the server only if it is sufficiently informative. When the latest gradient is not available, the stale one will be reused at the server. To implement this communication-censoring strategy, the batch sizes are increasing in order to alleviate the effect of gradient noise. Theoretically, CSGD enjoys the same order of convergence rate as that of SGD, but effectively reduces communication. Numerical experiments further demonstrate the sizable communication saving of CSGD.

Countering online hate speech is a critical yet challenging task, but one which can be aided by the use of Natural Language Processing (NLP) techniques. Previous research has primarily focused on the development of NLP methods to automatically and effectively detect online hate speech while disregarding further action needed to calm and discourage individuals from using hate speech in the future. In addition, most existing hate speech datasets treat each post as an isolated instance, ignoring the conversational context. In this paper, we propose a novel task of generative hate speech intervention, where the goal is to automatically generate responses to intervene during online conversations that contain hate speech. As a part of this work, we introduce two fully-labeled large-scale hate speech intervention datasets collected from Gab and Reddit. These datasets provide conversation segments, hate speech labels, as well as intervention responses written by Mechanical Turk Workers. In this paper, we also analyze the datasets to understand the common intervention strategies and explore the performance of common automatic response generation methods on these new datasets to provide a benchmark for future research.

Just about every estimate shows that the cannabis industry will see strong long-term growth. Yet there are some major challenges–and they are more than just about changing existing laws and regulations. But AI (Artificial Intelligence) is likely to be a big help. True, the industry has not been a big adopter of new technologies. However, this should change soon as investors pour billions of dollars into the space.

Kai-Fu Lee is a Chinese venture capitalist, technology executive, writer, and computer scientist. He is currently based in Beijing, China. Lee developed the world's first speaker-independent, continuous speech recognition system as his Ph.D. thesis at Carnegie Mellon. He later worked as an executive, first at Apple, then SGI, Microsoft, and then Google.He became the focus of a 2005 legal dispute between Google and Microsoft, his former employer, due to a one-year non-compete agreement that he signed with Microsoft in 2000 when he became its corporate vice president of interactive services.

What Y ou See Is What Y ou Get? Abstract Although systematic biases in decision-making are widely documented, the ways in which they emerge from different sources is less understood. We present a controlled experimental platform to study gender bias in hiring by decoupling the effect of world distribution (the gender breakdown of candidates in a specific profession) from bias in human decision-making. We explore the effectiveness of representation criteria, fixed proportional display of candidates, as an intervention strategy for mitigation of gender bias by conducting experiments measuring human decision-makers' rankings for who they would recommend as potential hires. Experiments across professions with varying gender proportions show that balancing gender representation in candidate slates can correct biases for some professions where the world distribution is skewed, although doing so has no impact on other professions where human persistent preferences are at play. We show that the gender of the decision-maker, complexity of the decision-making task and over-and under-representation of genders in the candidate slate can all impact the final decision. By decoupling sources of bias, we can better isolate strategies for bias mitigation in human-in-the-loop systems. Introduction Machine learning can aid decision-making and is used in recommendation systems that play increasingly prevalent roles in the world. We now deploy systems to help hire candidates (HireVue 2018), determine who to police more (V eale, V an Kleek, and Binns 2018), and assess the likelihood of an individual to recidivate on a crime (Angwin et al. 2016). Because these systems are trained on real world data, they often produce biased decision outcomes in a manner that is discriminatory against underrepresented groups. Systems have been found to unfairly discriminate against defendants of color in assessing bail (Angwin et al. 2016), incorrectly classify minority groups in facial recognition tasks (Raji and Buolamwini 2019), and engage in wage theft for honest workers (McInnis et al. 2016). A biased decision can be impacted by world, algorithmic, and human bias. Representation criteria is an intervention deployed when the candidate slate is generated.

If there was ever a time that ethics should be formally applied to technology, it is with the emergence of Artificial Intelligence. Yet most of the big AI companies struggle with what should seem a simple task: defining ethics for the use of their products. Without the underpinnings of a moral backbone, powerful tools often become a caustic capability for abuse. AI technology leaders must establish the guard-rails before chaos ensues. As the great strategist Sun Tzu professed "Plan for what is difficult while it is easy, do what is great while it is small".