Every day is Halloween for Virtual YouTubers or "VTubers" -- the new generation of wildly popular online entertainers whose voices and actions are represented in real time by colourful and expressive anime characters. Now, a Google researcher has released a deep neural network model that makes animating a VTube persona a little easier. Using motion capture systems to transfer human movements to cartoon characters in real-time is a process that can be traced back to the 90s. The approach however was not popularized, and the term "Virtual YouTuber" did not enter our vocabulary until the virtual character "Kizuna AI" debuted in 2016. Kizuna is a cute young girl with wide eyes and a pink butterfly bow perched atop her long flowing hair -- any otaku's dream.

In clinics around the world, a type of artificial intelligence called deep learning is starting to supplement or replace humans in common tasks such as analyzing medical images. Already, at Massachusetts General Hospital in Boston, "every one of the 50,000 screening mammograms we do every year is processed through our deep learning model, and that information is provided to the radiologist," says Constance Lehman, chief of the hospital's breast imaging division. In deep learning, a subset of a type of artificial intelligence called machine learning, computer models essentially teach themselves to make predictions from large sets of data. The raw power of the technology has improved dramatically in recent years, and it's now used in everything from medical diagnostics to online shopping to autonomous vehicles. But deep learning tools also raise worrying questions because they solve problems in ways that humans can't always follow.

If you're trying to solve data related problems with no or limited resources, be them time, money or skills, don't go no further. This talk is opinionated and deals with GDPR, deep learning, and all the hype. How does data infiltrate the organization? Which roles come first, what problems do they solve and what problems do they introduce? A down-to-earth approach in this hype-driven environment to make decisions impactful and practical-based on real world experience, not product brochures and GitHub repository stars.

The AWS Deep Learning AMIs are available on Ubuntu 18.04, Ubuntu 16.04, Amazon Linux 2, and Amazon Linux with TensorFlow 1.15, Tensorflow 2.0, PyTorch 1.3.1, Also new in this version is support for AWS Neuron, a SDK for running inference using AWS Inferentia chips. It consists of a compiler, run-time, and profiling tools that enable developers to run high-performance and low latency inference using Inferentia-based EC2 Inf1 instances. Neuron is pre-integrated into popular machine learning frameworks including TensorFlow, Pytorch, and MXNet to deliver optimal performance of EC2 Inf1 instances. Customers using Amazon EC2 Inf1 instances will receive the highest performance and lowest cost for machine learning inference in the cloud, and no longer need to make the sub-optimal tradeoff between optimizing for latency or throughput when running large machine learning models in production.

With AI, the future promises longer life expectancy, increased productivity, and better preservation of precious resources. You will be able to take a picture of a mole on your leg and send it electronically to a dermatologist, who will use deep neural networks to determine whether it is skin cancer. Data-driven sensors and drones will determine the perfect amount of pesticide and water to promote agricultural diversity and counter monocropping. The AI revolution in transportation will herald autonomous planes, trains and automobiles. Music will be created to improve not only mood but heart rate and brain activity.

This means that the test favors "program synthesis," the subfield of AI that involves generating programs that satisfy high-level specifications. This approach is in contrast with current trends in AI, which are inclined toward creating programs that are optimized for a limited set of tasks (e.g., playing a single game). In his experiments with ARC, Chollet has found that humans can fully solve ARC tests.

There's a lot of convention behind the term "artificial intelligence," and potentially that is the problem. Conventional models for AI, which are based on how the human brain might work, are not effective as we still don't have a definitive understanding of how the brain works, says Eberhard Schoeneburg, founder of Alternative AI. He believes a new way of thinking must be adapted for AI. "Even if you have a very simplified model of the brain, it wouldn't solve all these issues or all these problems. The key aspect of Alternative AI is to come up with explaining intelligence without referring to brains," says Schoeneburg. But quantum processes in nature can be studied for insights to create AI with actual intelligence, known as Artificial General Intelligence (AGI). That may soon become a reality. As Google claims "quantum supremacy" in the developing field of quantum computing, some experts suggest the breakthrough could be a boon to the field of artificial intelligence (AI) and vice-versa. In a recent interview with MIT Technology Review, Google CEO Sundar Pichai gave credence to AI as it "can accelerate quantum computing and quantum computing can accelerate AI." See related article: How blockchain can save A.I. Deep learning methods used in AI currently have narrow use cases which rely on static pattern recognition, while a quantum-based system may be more suited for real life applications, says Schoeneburg. Nonetheless, other analysts are less bullish on the prospect of quantum computing applications in the short term. Schoeneburg explains how artificial intelligence should adapt to quantum technology and more. This Forkast.News exclusive brings together two leading voices in artificial intelligence today: Susan Oh, founder of Muckr AI and who also serves as cochair of AI, Blockchain for Impact for the United Nations General Assembly, sits down with the "Godfather of Alternative AI" Eberhard Schoeneburg and calls out "deep learning" as being too specific to be "intelligent." To understand the future of AI, one must understand the roots of its past. Susan Oh: I have the great honor of sitting down with Eberhard Schoeneburg, who is the godfather of Alternative AI. He's also the man that gave us [one of the first] chatbots, though he says that he thinks it's a gimmick and bullsh*t now. So Eberhard, thank you so much for sitting down with me. I think we both agree that AI has failed to live up to the hype and the promise. I don't think what people realize is that this is the fifth wave of AI, that people have been working on intelligent computing systems since the 1950s.

During the Intel AI Summit earlier this month where the company demonstrated its initial processors for artificial intelligence training and inference workloads, Naveen Rao, corporate vice president and general manager of the Artificial Intelligence Products Group at Intel, spoke about the rapid pace of evolution in the AI space that also includes machine learning and deep learning. The Next Platform did an in-depth look at the technical details Rao shared about the products. But as noted in the story, Rao explained that the complexity of neural network models – when talking about the number of parameters – is growing ten-fold every year, a rate that is unlike any other technology trend we have ever seen. For Intel and the myriad other tech vendors getting making inroads into the space, AI and components like machine learning and deep learning already is a big business and promises to get bigger. Intel's AI products are expected to generate more than $3.5 billion in revenue for the chip maker this year, according to Rao.

The Radiological Society of North America (RSNA) is an international society of radiologists, medical physicists, and other medical professionals. The annual RSNA meeting with approximately 50,000 attendees in Chicago is a place for scientific exchange and clinical training. This year radiologists are invited to experience the hands-on cutting-edge technology of artificial intelligence, 3D printing, and virtual reality. The Fraunhofer MEVIS team will be pleased to welcome you at their booth, located at the "AI Showcase". Our experts are looking forward to providing you with a range of latest developments in deep learning and artificial intelligence, e.g., our free software MEVIS draw.

We study the phenomenon that some modules of deep neural networks (DNNs) are more critical than others. Meaning that rewinding their parameter values back to initialization, while keeping other modules fixed at the trained parameters, results in a large drop in the network's performance. Our analysis reveals interesting properties of the loss landscape which leads us to propose a complexity measure, called module criticality, based on the shape of the valleys that connects the initial and final values of the module parameters. We formulate how generalization relates to the module criticality, and show that this measure is able to explain the superior generalization performance of some architectures over others, whereas earlier measures fail to do so. 1 Introduction Neural networks have had tremendous practical impact in various domains such as revolutionizing many tasks in computer vision, speech and natural language processing. However, many aspects of their design and analysis have remained mysterious to this date. One of the most important questions is "what makes an architecture work better than others given a specific task?" Extensive research in this area has led to many potential explanations on why some types of architectures have better performance; however, we lack a unified view that provides a complete and satisfactory answer. In order to attain a unified view on superiority of one architecture over another in terms of generalization performance, we need to come up with a measure that effectively captures this. Analyzing the generalization behavior of neural networks has been an active area of research since Baum and Haussler (1989). Many generalization bounds and complexity measures have been proposed so far. Bartlett (1998) emphasized on the norm of the weights in predicting the generalization error.