Google ditched autopilot driving feature after user napped

Daily Mail - Science & tech

Waymo, a self-driving car company spun out from Google's parent company Alphabet, has stopped developing features that require drivers to take control in dangerous situations, its chief executive has said. Autopilot reliance has left users prone to distractions and ill-prepared to manoeuvre, the company said. The decision followed experiments of the technology in Silicon Valley that showed test users napping, putting on makeup and fiddling with their phones as the vehicles travelled up to 56mph (90kph). Alphabet's self-driving car unit Waymo (pictured) has stopped developing features that required drivers to take control in dangerous situations, its chief executive said on Monday Other self-driving automakers include similar autopilot features for highway-driving in vehicles, but they require drivers to take over the steering wheel in tricky situations. Waymo planned to do the same.


Exam Ref 70-741 Networking with Windows Server 2016 - Programmer Books

#artificialintelligence

Microsoft Exam 70-741 is for IT Pros interested in enhancing their skills for the latest version of Windows Server. This exam validates skills and knowledge for configuring advanced tasks required to deploy, manage, and maintain a Windows Server 2016 infrastructure, such as fault tolerance, certificate services, and identity federation. Passing this exam along with two other exams is required for MCSA and MCSE certifications. The Exam Ref is the official study guide for Microsoft certification exam 70-741. Featuring concise, objective-by-objective reviews and strategic case scenarios and Thought Experiments, exam candidates get professional-level preparation for the exam.


Could this be the start of a space federation? UN to launch its first mission into low-Earth orbit

Daily Mail - Science & tech

The International Space Station has been a beacon of what can be achieved when countries work together to venture out into our solar system. But now the United Nations is planning to enter the space race itself after signing a deal to launch a space plane into low Earth orbit in 2021. It is hoping to give those countries too poor to afford their own space programme the chance to send payloads into orbit around the Earth. The UN has signed a deal with Sierra Nevada to send one of its Dream Chaser space planes (pictured) on a 14-day mission into low Earth orbit in 2021. The UN's Office of Outer Space Affairs has signed a deal with private space firm Sierra Nevada to have the sole use of one of its Dream Chaser space planes.


Nobel laureate Tasuku Honjo hopes Japan invests more in science

The Japan Times

KYOTO – Nobel Prize-winning scientist Tasuku Honjo voiced hope on Tuesday that Japan would invest more in science, a day after he was chosen for this year's award in physiology or medicine along with American James Allison for their studies on cancer therapy. "I was able to prove that it is not rare for fundamental research to lead to applications," Honjo, 76, said at a news conference held at Kyoto University, where he is currently a professor. "Science is an investment for the future." News that Honjo became the 26th Japanese Nobel Prize winner was met with a shower of praise from cancer patient groups and the Japanese government on Monday. "Cancer patients are being saved by (the new cancer medicine) Opdivo, which originated from a study carried out by the Japanese researcher.


Model Pruning Enables Efficient Federated Learning on Edge Devices

arXiv.org Machine Learning

Federated learning is a recent approach for distributed model training without sharing the raw data of clients. It allows model training using the large amount of user data collected by edge and mobile devices, while preserving data privacy. A challenge in federated learning is that the devices usually have much lower computational power and communication bandwidth than machines in data centers. Training large-sized deep neural networks in such a federated setting can consume a large amount of time and resources. To overcome this challenge, we propose a method that integrates model pruning with federated learning in this paper, which includes initial model pruning at the server, further model pruning as part of the federated learning process, followed by the regular federated learning procedure. Our proposed approach can save the computation, communication, and storage costs compared to standard federated learning approaches. Extensive experiments on real edge devices validate the benefit of our proposed method.