To compile this list, we explored deep learning MOOCs (Massive Open Online Courses) published by top universities, colleges, and leading tech companies. Dedicated to beginners, intermediate, and advanced learners, and covering most concepts of Deep Learning, from the most basic to the cutting-edge, all of these courses are free and self-paced, and some of them even offer certificates. It goes without saying that all of these courses come with some prerequisites: basic knowledge of mathematics, how to manipulate GitHub repositories, and a good command of programming languages like Python. Google has published an online course dedicated to deep learning via Udacity, the online course platform. Google's MOOC trains intermediate to advanced developers free of charge for 12 weeks on many aspects of deep learning, such as how to build and optimize deep neural networks.

The next generation of AI applications will continuously interact with the environment and learn from these interactions. These applications impose new and demanding systems requirements, both in terms of performance and flexibility. In this paper, we consider these requirements and present Ray---a distributed system to address them. Ray implements a dynamic task graph computation model that supports both the task-parallel and the actor programming models. To meet the performance requirements of AI applications, we propose an architecture that logically centralizes the system's control state using a sharded storage system and a novel bottom-up distributed scheduler. In our experiments, we demonstrate sub-millisecond remote task latencies and linear throughput scaling beyond 1.8 million tasks per second. We empirically validate that Ray speeds up challenging benchmarks and serves as both a natural and performant fit for an emerging class of reinforcement learning applications and algorithms.

Recently, deep neural networks (DNNs) have been regarded as the state-of-the-art classification methods in a wide range of applications, especially in image classification. Despite the success, the huge number of parameters blocks its deployment to situations with light computing resources. Researchers resort to the redundancy in the weights of DNNs and attempt to find how fewer parameters can be chosen while preserving the accuracy at the same time. Although several promising results have been shown along this research line, most existing methods either fail to significantly compress a well-trained deep network or require a heavy fine-tuning process for the compressed network to regain the original performance. In this paper, we propose the \textit{Block Term} networks (BT-nets) in which the commonly used fully-connected layers (FC-layers) are replaced with block term layers (BT-layers). In BT-layers, the inputs and the outputs are reshaped into two low-dimensional high-order tensors, then block-term decomposition is applied as tensor operators to connect them. We conduct extensive experiments on benchmark datasets to demonstrate that BT-layers can achieve a very large compression ratio on the number of parameters while preserving the representation power of the original FC-layers as much as possible. Specifically, we can get a higher performance while requiring fewer parameters compared with the tensor train method.

In Japan, famous for the longevity of its people, their endearing use of engrish, and their fetish for girls in Catholic school uniforms, the care of the elderly is a particularly acute problem. A third of the Japanese population is reportedly above the age of 60, and the number of people over 90 years of age just topped two million for the first time. Add in a rapidly shrinking population, and you have a country where you have more people eating the early bird special than not. So it's no surprise that Japan is leading the world in robotic elder care, offering a glimpse into our geriatric future.

Unlike tennis balls and musketeers, AI ventures don't typically come in threes. The former AI leader at both Google and Baidu has been on an entrepreneurial spree, making his third major announcement in recent months to launch Landing.ai. The startup will help make large manufacturing companies more efficient by using artificial intelligence, according Ng's blog post today (Dec. He also serves as chairman of Coursera, the online education company he cofounded in 2012. Ng writes that his startup and Foxconn will "jointly develop and deploy AI solutions and training globally."

Gargantuan Taiwanese manufacturer Foxconn employs more than 1 million people and tens of thousands of robots making iPhones and other electronics. It has a reputation for cost cutting, including at the expense of its workers. Now, it's teaming up with an artificial-intelligence researcher who helped trigger Google's reorientation around machine learning in order to make its own factories more efficient. Andrew Ng was a Stanford professor when he joined Google in 2011 to work on a project that created software able to recognize cats--and a new corporate emphasis on AI at Google. He later led AI research at Chinese search engine Baidu.

AI is coming for your job. AI is taking over the world. If we compiled all the headlines about artificial intelligence from the last year, we'd have a picture of a dystopian world where jobs are scarce and AI and automation rule everything we do. In this scenario, millions of people are impacted by AI and autonomous systems created with little regard for their consequences: They are deployed in unethical ways, riddled with errors and bias, and discriminatory. The obscurity of how AI works and where it's used result in fear and confusion.

"Deep Learning" computer systems, based on artificial neural networks that mimic the way the brain learns from an accumulation of examples, have become a hot topic in computer science. In addition to enabling technologies such as face- and voice-recognition software, these systems could scour vast amounts of medical data to find patterns that could be useful diagnostically, or scan chemical formulas for possible new pharmaceuticals. But the computations these systems must carry out are highly complex and demanding, even for the most powerful computers. Now, a team of researchers at MIT and elsewhere has developed a new approach to such computations, using light instead of electricity, which they say could vastly improve the speed and efficiency of certain deep learning computations. Their results appear today in the journal Nature Photonics in a paper by MIT postdoc Yichen Shen, graduate student Nicholas Harris, professors Marin Soljacic and Dirk Englund, and eight others.

Artificial Intelligence is transforming the landscape of human evolution, with one of the biggest landmarks for the technology being the declaration of an AI robot - Sophia - as a "national citizen" of the Kingdom of Saudi Arabia. Despite the fact that the earliest footprints of Artificial Intelligence can be traced back to 1956, its impact had been behind the curtains, until recently. During the last decade, AI has sparked polar reactions from observers, with some crowning it as the "face of the future", while others demarking it as the "beginning of the end". Elon Musk, the CEO of Tesla Motors, went as far as saying that AI would largely be attributable for the outburst of World War-III. On the other side of the fence, however, we have the likes of Mark Zuckerberg and Bill Gates, who could not have expressed more hope and faith in the revelation.

If there's any emerging technology that's gripped the public consciousness in recent years it's AI and machine learning (ML). Autonomous vehicles, shopping recommendations, Siri and Alexa, these are just a few of the day to day examples of the rapid evolution of ML applications. The fervour around AI and ML's development is only fuelling these advancements. As public interest grows we're already seeing more students attracted to ML and AI courses. Just look at the popularity of Professor Andrew Ng's Coursera course on machine learning or the record number of Stanford students who enrolled in the machine learning class this semester.