Google, a worldwide leader in artificial intelligence and machine learning, has eagerly expanded its educational and professional resources needed to make AI and ML more accessible. The efforts, however, are not purely altruistic. By breeding familiarity with its TensorFlow software, Google is ensuring that the next generation of AI and ML experts are familiar with its platform and tools. The expansion of the AI workforce is certainly good for the overall market, but Google is still making an investment in its own future strategy. The breadth and depth of Googlers' expertise, especially in AI and ML, make them prime targets for poaching by other organizations looking to build out their own advanced technology stack.

Sophie Turol is passionate about delivering well-structured articles that cater for picky technical audience. With 3 years in technical writing and 5 years in editorship, she enjoys collaboration with developers to create insightful, yet intelligible technical tutorials, overviews, and case studies. Sophie is enthusiastic about deep learning solutions--TensorFlow in particular--and PaaS systems, such as Cloud Foundry.

When we look at a stack of blocks or a stack of Oreos, we intuitively have a sense of how stable it is, whether it might fall over, and in what direction it may fall. That's a fairly sophisticated calculation involving the mass, texture, size, shape, and orientation of the objects in the stack. Researchers at MIT led by Josh Tenenbaum hypothesize that our brains have what you might call an intuitive physics engine: The information that we are able to gather through our senses is imprecise and noisy, but we nonetheless make an inference about what we think will probably happen, so we can get out of the way or rush to keep a bag of rice from falling over or cover our ears. Such a "noisy Newtonian" system involves probabilistic understandings and can fail. Consider this image of rocks stacked in precarious formations.

One year back, I was a newbie to the world of Machine Learning. I used to get overwhelmed by small decisions, like choosing the language to code with, choosing the right online courses, or choosing the correct algorithms. So, I have planned to make it easier for folks to get into Machine Learning. I'll assume that many of us are starting from scratch on our Machine Learning journey. Let's find out how current professionals in the field reached their destination, and how we can emulate them on our journey. I will illustrate how you can learn Data Science by drawing a parallel between how Rafael Nadal learned to play tennis, and how you can learn Machine Learning.

During college, while doing a geophysics internship aboard an oil rig, I realized that software was the future--so I switched my major to computer science. After more than a decade working at Google, I had a similar moment where I realized that AI is the future of computer science. Today, I lead Google's machine learning education effort, in the hope of making AI and its benefits accessible to everyone. AI can solve complex problems and has the potential to transform entire industries, which means it's crucial that AI reflect a diverse range of human perspectives and needs. That's why part of Google AI's mission is to help anyone interested in machine learning succeed--from researchers, to developers and companies, to students like Abu.

Machine learning and AI are some of the biggest topics in the tech world right now, and Google is looking to make those fields more accessible to more people with its new Learn with Google AI website. Google has been pursuing AI education for a while, both with advanced projects like TensorFlow and more playful projects like cat doodles and a machine vision experiment meant to showcase AI projects in more practical ways. Google envisions the Learn with Google AI site serving as a repository for machine learning and AI, and it's meant to be a hub for anyone looking to "learn about core ML concepts, develop and hone your ML skills, and apply ML to real-world problems." The site will apparently cater to all levels of AI enthusiasts, from researchers looking for advanced tutorials to beginners. The site also features a free course called Machine Learning Crash Course (MLCC).

Generative adversarial networks (GAN) have been effective for learning generative models for real-world data. However, existing GANs (GAN and its variants) tend to suffer from training problems such as instability and mode collapse. In this paper, we propose a novel GAN framework called evolutionary generative adversarial networks (E-GAN) for stable GAN training and improved generative performance. Unlike existing GANs, which employ a pre-defined adversarial objective function alternately training a generator and a discriminator, we utilize different adversarial training objectives as mutation operations and evolve a population of generators to adapt to the environment (i.e., the discriminator). We also utilize an evaluation mechanism to measure the quality and diversity of generated samples, such that only well-performing generator(s) are preserved and used for further training. In this way, E-GAN overcomes the limitations of an individual adversarial training objective and always preserves the best offspring, contributing to progress in and the success of GANs. Experiments on several datasets demonstrate that E-GAN achieves convincing generative performance and reduces the training problems inherent in existing GANs.

Online structure learning approaches, such as those stemming from Statistical Relational Learning, enable the discovery of complex relations in noisy data streams. However, these methods assume the existence of fully-labelled training data, which is unrealistic for most real-world applications. We present a novel approach for completing the supervision of a semi-supervised structure learning task. We incorporate graph cut minimisation, a technique that derives labels for unlabelled data, based on their distance to their labelled counterparts. In order to adapt graph cut minimisation to first order logic, we employ a suitable structural distance for measuring the distance between sets of logical atoms. The labelling process is achieved online (single-pass) by means of a caching mechanism and the Hoeffding bound, a statistical tool to approximate globally-optimal decisions from locally-optimal ones. We evaluate our approach on the task of composite event recognition by using a benchmark dataset for human activity recognition, as well as a real dataset for maritime monitoring. The evaluation suggests that our approach can effectively complete the missing labels and eventually, improve the accuracy of the underlying structure learning system.

How does machine learning differ from traditional programming? What is loss, and how do I measure it? How does gradient descent work? How do I determine whether my model is effective? How do I represent my data so that a program can learn from it? How do I build a deep neural network?

Some of us understand how Artificial Intelligence will impact manufacturing or R&D, but what about other areas of the business? For example, what role will it play in Learning & Development? What do leaders in L&D and HR need to consider in developing, using and promoting the use of AI products to their internal customers? Can it be used effectively to teach management skills? How is bias eliminated in such a program?