In stochastic optimization, large batch training can leverage parallel resources to produce faster wall-clock training times per epoch. However, for both training loss and testing error, recent results analyzing large batch Stochastic Gradient Descent (SGD) have found sharp diminishing returns beyond a certain critical batch size. In the hopes of addressing this, the Kronecker-Factored Approximate Curvature (\mbox{K-FAC}) method has been hypothesized to allow for greater scalability to large batch sizes for non-convex machine learning problems, as well as greater robustness to variation in hyperparameters. Here, we perform a detailed empirical analysis of these two hypotheses, evaluating performance in terms of both wall-clock time and aggregate computational cost. Our main results are twofold: first, we find that \mbox{K-FAC} does not exhibit improved large-batch scalability behavior, as compared to SGD; and second, we find that \mbox{K-FAC}, in addition to requiring more hyperparameters to tune, suffers from the same hyperparameter sensitivity patterns as SGD. We discuss extensive results using residual networks on \mbox{CIFAR-10}, as well as more general implications of our findings.

Online learning has traditionally focused on the expected rewards. In this paper, a risk-averse online learning problem under the performance measure of the mean-variance of the rewards is studied. Both the bandit and full information settings are considered. The performance of several existing policies is analyzed, and new fundamental limitations on risk-averse learning is established. In particular, it is shown that although a logarithmic distribution-dependent regret in time $T$ is achievable (similar to the risk-neutral problem), the worst-case (i.e. minimax) regret is lower bounded by $\Omega(T)$ (in contrast to the $\Omega(\sqrt{T})$ lower bound in the risk-neutral problem). This sharp difference from the risk-neutral counterpart is caused by the the variance in the player's decisions, which, while absent in the regret under the expected reward criterion, contributes to excess mean-variance due to the non-linearity of this risk measure. The role of the decision variance in regret performance reflects a risk-averse player's desire for robust decisions and outcomes.

How to obtain a model with good interpretability and performance has always been an important research topic. In this paper, we propose rectified decision trees (ReDT), a knowledge distillation based decision trees rectification with high interpretability, small model size, and empirical soundness. Specifically, we extend the impurity calculation and the pure ending condition of the classical decision tree to propose a decision tree extension that allows the use of soft labels generated by a well-trained teacher model in training and prediction process. It is worth noting that for the acquisition of soft labels, we propose a new multiple cross-validation based method to reduce the effects of randomness and overfitting. These approaches ensure that ReDT retains excellent interpretability and even achieves fewer nodes than the decision tree in the aspect of compression while having relatively good performance. Besides, in contrast to traditional knowledge distillation, back propagation of the student model is not necessarily required in ReDT, which is an attempt of a new knowledge distillation approach. Extensive experiments are conducted, which demonstrates the superiority of ReDT in interpretability, compression, and empirical soundness.

Some involve coding, some are drag-and-drop, some are difficult for beginners, some have no coding at all. All of these tools will help you with data visualization. But one of the most overlooked but critical practical functions of a data scientist has been included under this heading: summarisation. Summarisation means the practical result of your data science workflow. What does the result of your analysis mean for the operation of the business or the research problem that you are currently working on? How do you convert your result to the maximum improvement for your business? Can you measure the impact this result will have on the profit of your enterprise?

As Artificial Intelligence technology continues to grow, you may be wondering, "how will AI impact the future of business, jobs, and daily life?". If you're a student (or know one), you can turn those thoughts & ideas into a scholarship opportunity! Currently enrolled college students are eligible for the Bold360 Artificial Intelligence Scholarship program!

BUSINESSES in Australia such as law and accounting firms, technology companies, and medical facilities are staffed with professionals certified by government bodies. In order to ensure these professionals stay up-to-date and relevant, the governing bodies often require that they receive training on an ongoing basis. CPA Australia and the Lawyers Society of South Australia, for example, require members undergo 20 and 10 hours of CPD training per year and offer seminars and sessions to help meet that requirement. However, practically speaking, the training on offer might not be directly relevant to the businesses or jobs that these professionals are performing on a daily basis. For example, CPA Australia might offer a seminar on understanding wealth management in the accounting context. Although that knowledge is relevant to a CPA in general, it might not be suited to someone handling internal audit for a manufacturing entity.

Microsoft has launched the AI Business School, an online series of case studies and free instructional videos made to help business executives design and successfully implement an AI strategy within their organization. The Microsoft AI Business School follows the launch of an AI school for developers and AI School last year. AI Business School is born out of three years of conversations with customers implementing AI, as well as lessons learned from AI solutions Microsoft introduced internally, says Mitra Azizirad, Microsoft vice president of AI marketing and productization. Course content will focus on four main areas: strategy, culture, technology basics, and responsible AI. And courses will include tools for evaluating a business' AI maturity level to understand what's required to successfully implement AI, for example.

Artificial Intelligence is nowadays one of the most popular topics to discus upon. It only continues to significantly increase in its role in business and daily life in general. It is important to discuss that what role women are likely to play in the world of Artificial Intelligence, writes Nandita Koshal, Research Associate, International Institute for Higher Education Research and Capacity Building, O P Jindal Global University. AI and IoT technologies are increasingly being fashioned on women and their perceived roles. They play the role of our guide as GPS that takes us to our destination; as Cortana and Siri they become our personal assistants who aid in expediting our day to day commitments; and as Alexa and Google play they become our companions and friends who respect and fulfil our requests.

We organized a competition on Autonomous Lifelong Machine Learning with Drift that was part of the competition program of NeurIPS 2018. This data driven competition asked participants to develop computer programs capable of solving supervised learning problems where the i.i.d. assumption did not hold. Large data sets were arranged in a lifelong learning and evaluation scenario and CodaLab was used as the challenge platform. The challenge attracted more than 300 participants in its two month duration. This chapter describes the design of the challenge and summarizes its main results.

One of the main challenges of advancing task-oriented learning such as visual task planning and reinforcement learning is the lack of realistic and standardized environments for training and testing AI agents. Previously, researchers often relied on ad-hoc lab environments. There have been recent advances in virtual systems built with 3D physics engines and photo-realistic rendering for indoor and outdoor environments, but the embodied agents in those systems can only conduct simple interactions with the world (e.g., walking around, moving objects, etc.). Most of the existing systems also do not allow human participation in their simulated environments. In this work, we design and implement a virtual reality (VR) system, VRKitchen, with integrated functions which i) enable embodied agents powered by modern AI methods (e.g., planning, reinforcement learning, etc.) to perform complex tasks involving a wide range of fine-grained object manipulations in a realistic environment, and ii) allow human teachers to perform demonstrations to train agents (i.e., learning from demonstration). We also provide standardized evaluation benchmarks and data collection tools to facilitate a broad use in research on task-oriented learning and beyond.