The model reduction problem that eases the computation costs and latency of complex deep learning architectures has received an increasing number of investigations owing to its importance in model deployment. One promising method is knowledge distillation (KD), which creates a fast-to-execute student model to mimic a large teacher network. In this paper, we propose a method, called KDFM (Knowledge Distillation with Feature Maps), which improves the effectiveness of KD by learning the feature maps from the teacher network. Two major techniques used in KDFM are shared classifier and generative adversarial network. Experimental results show that KDFM can use a four layers CNN to mimic DenseNet-40 and use MobileNet to mimic DenseNet-100. Both student networks have less than 1\% accuracy loss comparing to their teacher models for CIFAR-100 datasets. The student networks are 2-6 times faster than their teacher models for inference, and the model size of MobileNet is less than half of DenseNet-100's.

Quantitative definitions of what is unfair and what is fair have been introduced in multiple disciplines for well over 50 years, including in education, hiring, and machine learning. We trace how the notion of fairness has been defined within the testing communities of education and hiring over the past half century, exploring the cultural and social context in which different fairness definitions have emerged. In some cases, earlier definitions of fairness are similar or identical to definitions of fairness in current machine learning research, and foreshadow current formal work. In other cases, insights into what fairness means and how to measure it have largely gone overlooked. We compare past and current notions of fairness along several dimensions, including the fairness criteria, the focus of the criteria (e.g., a test, a model, or its use), the relationship of fairness to individuals, groups, and subgroups, and the mathematical method for measuring fairness (e.g., classification, regression). This work points the way towards future research and measurement of (un)fairness that builds from our modern understanding of fairness while incorporating insights from the past.

Tencent is pushing ahead with efforts to sign up schools to use its all-in-one WeChat app as part of a broader shift by the company to go beyond consumers and serve more institutions. Since WeChat started partnering with public schools and private training centres in 2014, students have been to make purchases at canteens, check test scores and communicate in group chats with their peers, parents and teachers using the app's multiple functions. Tencent revamped its organisation structure on the eve of its 20th anniversary this year to align the company to compete for the industrial internet. In an open letter last month, chairman Pony Ma Huateng laid out the rationale for the restructuring, saying that digitisation of the economy meant the battle for the internet has moved on from consumers to industry. "WeChat Pay has been a tool to connect people, and is now increasingly connecting industries under the company's new direction to embrace the industrial internet."

Applying artificial intelligence to self-driving cars to smooth traffic, reduce fuel consumption, and improve air quality predictions may sound like the stuff of science fiction, but researchers at the Department of Energy's Lawrence Berkeley National Laboratory (Berkeley Lab) have launched two research projects to do just that. In collaboration with UC Berkeley, Berkeley Lab scientists are using deep reinforcement learning, a computational tool for training controllers, to make transportation more sustainable. One project uses deep reinforcement learning to train autonomous vehicles to drive in ways to simultaneously improve traffic flow and reduce energy consumption. A second uses deep learning algorithms to analyze satellite images combined with traffic information from cell phones and data already being collected by environmental sensors to improve air quality predictions. "Thirty percent of energy use in the U.S. is to transport people and goods, and this energy consumption contributes to air pollution, including approximately half of all nitrogen oxide emissions, a precursor to particular matter and ozone – and black carbon (soot) emissions," said Tom Kirchstetter, director of Berkeley Lab's Energy Analysis and Environmental Impacts Division, an adjunct professor at UC Berkeley, and a member of the research team.

AI is not a silver bullet, but it could help tackle some of the world's most challenging social problems. Artificial intelligence (AI) has the potential to help tackle some of the world's most challenging social problems. To analyze potential applications for social good, we compiled a library of about 160 AI social-impact use cases. They suggest that existing capabilities could contribute to tackling cases across all 17 of the UN's sustainable-development goals, potentially helping hundreds of millions of people in both advanced and emerging countries. Real-life examples of AI are already being applied in about one-third of these use cases, albeit in relatively small tests. They range from diagnosing cancer to helping blind people navigate their surroundings, identifying victims of online sexual exploitation, and aiding disaster-relief efforts (such as the flooding that followed Hurricane Harvey in 2017). AI is only part of a much broader tool kit of measures that can be used to tackle societal issues, however. For now, issues such as data accessibility and shortages of AI talent constrain its application for social good. This article is a condensed version of our discussion paper, Notes from the AI frontier: Applying AI for social good (PDF–3MB).

Cutting Edge Deep Learning For Coders, Part 2 from fast.ai

In the past few years, there's been a lot of discussion around Artificial Intelligence and it's implications for everything from industrial applications to smart phone apps--and if you have a smart phone, chances are you've come across TED Talks. If you haven't listened to a TED Talk, you should. Presented by people who are deeply connected to the field they are speaking about, it gives you the opportunity to listen to people with an insider's expertise that most of us would never hear from, but who can provide valuable perspectives. So whether or not you've ever seen a TED Talk, if you've been looking to know more about the current state and future direction of Artificial Intelligence in our world, these 10 TED Talks are essential presentations to help you become fully informed about the state of this revolutionary technology. Presented by Educator and Entrepreneur Sebastian Thrun & TED Conference Curator Chris Anderson, this Ted Talk is an excellent overview of what people talk about when they talk about artificial intelligence, what people are concerned about, and why.

"Machine learning" and "data science" - two terms that might sound mysterious to someone who has never dealt with these processes before. However, if you have heard about machine learning before, you probably know that it is a process that aims to change the very fundaments of how our society is built... Sounds dramatic? Well, perhaps, but I do have a strong case for that - you'll find it in my machine learning course! Why Choose this Machine Learning Course? Machine learning, as a concept, is indeed a very interesting one.

People in the tech world love to say that everybody should learn to code -- and there are a lot of good reasons to pick up the trade: earn a higher salary, understand the implications of A.I. and cybersecurity, protect yourself from automation. While wages have been stagnant, there has been consistent growth of high-paying jobs in software (for now). Despite all the incentives bringing people into programming and other technology fields, we actually need more people willing to break out of it. Don't get me wrong: I teach computer science and computer programming. I believe wholeheartedly in the power of technology to improve lives.

One of the most difficult things about learning a new skill is finding time to study. Being able to complete assignments in between meetings or while traveling can make all the difference in the ability to make regular progress. Unfortunately, none of the online courses in programming, data science and machine learning I've taken over this past year have great mobile solutions. Much of the work still requires a laptop. After a great deal of searching, I finally found a solution in two applications that allow users to run Jupyter notebooks and python terminal commands, both of which are common tools for completing machine learning tasks.