We often hear about the ever-changing world of technology and how it impacts the practice of law and legal compliance. This is also significantly affects what individuals are entering the law practice, as well as their interests, skills, and desired job opportunities. I recently talked with an aspiring in-house lawyer, Albert J. Higgins, who is about to graduate from the Sandra Day O'Connor College of Law at Arizona State University (ASU), which is ranked 25th by US News and World Report for best law schools and 22nd globally by the Academic Ranking of World Universities. Albert's interest is in legal analytics, a field that many of us do not understand or comprehend how it is useful in the practice of law. K: Albert, tell me a little about yourself and your background.

Peer review, evaluation, and selection is a fundamental aspect of modern science. Funding bodies the world over employ experts to review and select the best proposals of those submitted for funding. The problem of peer selection, however, is much more general: a professional society may want to give a subset of its members awards based on the opinions of all members; an instructor for a MOOC or online course may want to crowdsource grading; or a marketing company may select ideas from group brainstorming sessions based on peer evaluation. We make three fundamental contributions to the study of procedures or mechanisms for peer selection, a specific type of group decision-making problem, studied in computer science, economics, and political science. First, we propose a novel mechanism that is strategyproof, i.e., agents cannot benefit by reporting insincere valuations. Second, we demonstrate the effectiveness of our mechanism by a comprehensive simulation-based comparison with a suite of mechanisms found in the literature. Finally, our mechanism employs a randomized rounding technique that is of independent interest, as it solves the apportionment problem that arises in various settings where discrete resources such as parliamentary representation slots need to be divided proportionally.

When using stochastic gradient descent to solve large-scale machine learning problems, a common practice of data processing is to shuffle the training data, partition the data across multiple machines if needed, and then perform several epochs of training on the re-shuffled (either locally or globally) data. The above procedure makes the instances used to compute the gradients no longer independently sampled from the training data set. Then does the distributed SGD method have desirable convergence properties in this practical situation? In this paper, we give answers to this question. First, we give a mathematical formulation for the practical data processing procedure in distributed machine learning, which we call data partition with global/local shuffling. We observe that global shuffling is equivalent to without-replacement sampling if the shuffling operations are independent. We prove that SGD with global shuffling has convergence guarantee in both convex and non-convex cases. An interesting finding is that, the non-convex tasks like deep learning are more suitable to apply shuffling comparing to the convex tasks. Second, we conduct the convergence analysis for SGD with local shuffling. The convergence rate for local shuffling is slower than that for global shuffling, since it will lose some information if there's no communication between partitioned data. Finally, we consider the situation when the permutation after shuffling is not uniformly distributed (insufficient shuffling), and discuss the condition under which this insufficiency will not influence the convergence rate. Our theoretical results provide important insights to large-scale machine learning, especially in the selection of data processing methods in order to achieve faster convergence and good speedup. Our theoretical findings are verified by extensive experiments on logistic regression and deep neural networks.

Today, those looking for a non-traditional education have limited access to online classrooms, especially ones that are for-credit and affordable. But Thomas Frey predicts that, within 14 years, learning from robots will be entirely commonplace -- even for children. Frey is a futurist who began as an engineer at IBM and went on to found the DaVinci Institute, a networking firm and think tank for technical innovation to bring about a brighter future. Frey gives lectures and interviews on strategies for progress to high-profile audiences at places like NASA, the New York Times, and various Fortune 500 companies. He told Business Insider that he sees a future where innovators will enhance and improve the current landscape of online education.

Artificial Intelligence has the potential to greatly improve and change education systems across the world. There is a strong possibility for artificial intelligence to be used to help teachers effectively streamline their instruction process and to help students receive much more personalized help that is specifically suited to their strengths and weaknesses. Ideally AI will also help to complete some of the more menial tasks that teachers and teaching assistants have to work on, freeing them up to spend even more time helping students. There has been a lot of progress in this field in the past few years, as more and more companies have been involved in projects that aim to augment, improve, and change the way teaching is done. The field of Education is definitely ripe for innovation, and the advancement of artificial intelligence may be able to provide that innovation.

Artificial Intelligence (AI) is the next big thing, but how can you use it to create a better learning experience? Look at any list of disruptive technologies and AI is likely to be at the top, it's the latest tech buzzword to take over the news media. As an L&D professional, what should you be paying close attention to? And what's safe to ignore? Our round up of the more useful applications of AI for learning includes many of the leading examples from academia and adult education that will be filtering their way into the workplace learning environment as AI becomes more widespread.

Deep reinforcement learning is poised to revolutionise the field of AI and represents a step towards building autonomous systems with a higher level understanding of the visual world. Currently, deep learning is enabling reinforcement learning to scale to problems that were previously intractable, such as learning to play video games directly from pixels. Deep reinforcement learning algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world. In this survey, we begin with an introduction to the general field of reinforcement learning, then progress to the main streams of value-based and policy-based methods. Our survey will cover central algorithms in deep reinforcement learning, including the deep $Q$-network, trust region policy optimisation, and asynchronous advantage actor-critic. In parallel, we highlight the unique advantages of deep neural networks, focusing on visual understanding via reinforcement learning. To conclude, we describe several current areas of research within the field.

In the age of machine learning and robotics, it is not only the future of work that is in the balance. The very meaning of a career must be reimagined. I have been amazed by the number and variety of people concerned with this issue. Old friends discuss it over dinner, it's one of the first topics broached with strangers on planes, and executives raise it in the context of both their workforce and their family. Two groups feel most affected: those in the middle of their career whose work will be made redundant and for whom retraining and finding new work will be difficult.

Finally, the main aim of this blog post is to give a well-intentioned advice about the importance of Mathematics in Machine Learning and the necessary topics and useful resources for a mastery of these topics. However, some Machine Learning enthusiasts are novice in Maths and will probably find this post disheartening (seriously, this is not my aim). For beginners, you don't need a lot of Mathematics to start doing Machine Learning. The fundamental prerequisite is data analysis as described in this blog post and you can learn the maths on the go as you master more techniques and algorithms. This entry was originally published on my LinkedIn page.

Scientists who build artificial intelligence and autonomous systems need a strong ethical understanding of the impact their work could have. More than 100 technology pioneers recently published an open letter to the United Nations on the topic of lethal autonomous weapons, or "killer robots". These people, including the entrepreneur Elon Musk and the founders of several robotics companies, are part of an effort that began in 2015. The original letter called for an end to an arms race that it claimed could be the "third revolution in warfare, after gunpowder and nuclear arms". The UN has a role to play, but responsibility for the future of these systems also needs to begin in the lab. The education system that trains our AI researchers needs to school them in ethics as well as coding.