John Oliver addressed bitcoin and the crypto-craze on Last Week Tonight yesterday evening, describing the phenomenon as, "Everything you don't understand about money combined with everything you don't understand about computers." The British comedian satirised consumers' rush to enter the digicoin market without fully understanding how it works or what much of its accompanying jargon actually means. "The vast majority of people buying these coins are not paying much attention to the details of the startups they are attached to, they are responding to the huge fervour," Oliver observed. "Many people are buying coins for no reason other than other people are buying them." Elsewhere in an extensive 25-minute address, Oliver sent up the "cult-like devotion" with which investors follow sub-Reddits and YouTubers preaching the bitcoin gospel.

Tech giant Google has now introduced a new easy-to-learn platform called'Learn with Google AI', which are a set of educational resources developed by Machine Learning experts at the company. This platform will help people learn about concepts, develop skills and apply artificial intelligence to problems in real life. The company mentioned in a blog, "To help everyone understand how AI can solve challenging problems, we've created a resource called Learn with Google AI. This site provides ways to learn about core ML concepts, develop and hone your ML skills, and apply ML to real-world problems. From deep learning experts looking for advanced tutorials and materials on TensorFlow, to "curious cats" who want to take their first steps with AI, anyone looking for educational content from ML experts at Google can find it here."

Muhammad Zafar Iqbal, a celebrated Bangladeshi author and academic, was stabbed in the back on March 3, while attending a programme at a university in the northeastern district of Sylhet. The lone attacker, Foyzur Rahman, was caught before he could continue stabbing Iqbal from behind. The 24-year-old attacker later said he wanted to kill Iqbal because he believed the academic was an "an enemy of Islam". Iqbal, who has authored more than 200 books, is now out of danger and is recuperating at the Combined Military Hospital (CMH) in the capital, Dhaka, according to his wife, Yasmin Haque. The stabbing has brought back memories of a series of fatal attacks against secular bloggers in Bangladesh nearly three years ago.

"This acquisition of ScanEagle is one way to modernize the military to deter those who want to wage war against our country," Defense Secretary Delfin Lorenzana said at a ceremony at a Manila air force base, attended by the U.S. ambassador.

Ever since its failure was confirmed in late 2015, China's Tiangong-1 spacecraft has been tracked by organizations to try determine where it will fall following reentry, expected to take place in the first week of April. While the loss of the module may have been a setback, it hasn't stopped China from pursuing its ambitious goal of building its own space station, Tiangong. Translating to "heavenly palace" from Mandarin, the Tiangong program is a bold one, considering China was a relative newcomer to the space race but has already established a habitable presence in orbit around Earth. Tiangong-1, before it failed, saw two visits by crews who spent time in the module, proving that it was habitable. The Chinese are following the tested approach of building a space station one module at a time.

When it comes to artificial intelligence, a lot of attention has been focused on issues of privacy and economics – what happens if AI makes human workers obsolete. Now, a new report from the non-profit Environmental Law Institute highlights the potential environmental impacts of AI-driven technologies, from autonomous cars to smart thermostats. Lead author Dave Rejeski says that whether those impacts are positive or negative will depend on how the technology is built and used, and the time to start thinking about that is now. The PyeongChang Olympics are likely to be remembered for the joint Korean team, wind delays, and robots. South Korea is taking advantage of the international spotlight to show off its leadership in robotics, with eleven different types of robots – eighty five, in all – in action at the Olympics.

Within the few months of its release, Internet of Things (IoT) has taken everyone by storm in numerous ways. As a result, more and more organizations, industries, and technologists catch the IoT bug. Right from Data streaming to data collection, events, decisions, processes, messaging, and integration, everything is involved in the form of developer's activities. Now, do you think that IoT is just another opportunity for vendors to sell or update developer tools? Well, according to several resources nearly 40 billion which is approximately 30 devices for each and every active social network user in the world.

Five simple soft sensor methodologies with two update conditions were compared on two experimentally-obtained datasets and one simulated dataset. The soft sensors investigated were moving window partial least squares regression (and a recursive variant), moving window random forest regression, the mean moving window of y, and a novel random forest partial least squares regression ensemble (RF-PLS), all of which can be used with small sample sizes so that they can be rapidly placed online. It was found that, on two of the datasets studied, small window sizes led to the lowest prediction errors for all of the moving window methods studied. On the majority of datasets studied, the RF-PLS calibration method offered the lowest onestep-ahead prediction errors compared to those of the other methods, and it demonstrated greater predictive stability at larger time delays than moving window PLS alone. It was found that both the random forest and RF-PLS methods most adequately modeled the datasets that did not feature purely monotonic increases in property values, but that both methods performed more poorly than moving window PLS models on one dataset with purely monotonic property values. Other data dependent findings are presented and discussed. Preprint submitted to Arxiv March 14, 2018 1. Introduction Soft sensors for regression tasks have found wide utility in process engineering and process analytical chemistry [1, 2, 3]. A soft sensor is effectively a calibration used on time-series data. Here, we consider a soft sensor to be any algorithm that can be used to estimate a property value from several readily available but indirect measurements. The goal of implementing a soft sensor is typically to avoid the use of a physical sensor for variables that may require extensive time or work up to measure [3]. In the context of industrial chemical processes, these algorithms should meet several specifications.

Uplift modeling is aimed at estimating the incremental impact of an action on an individual's behavior, which is useful in various application domains such as targeted marketing (advertisement campaigns) and personalized medicine (medical treatments). Conventional methods of uplift modeling require every instance to be jointly equipped with two types of labels: the taken action and its outcome. However, obtaining two labels for each instance at the same time is difficult or expensive in many real-world problems. In this paper, we propose a novel method of uplift modeling that is applicable to a more practical setting where only one type of labels is available for each instance. We demonstrate the effectiveness of the proposed method through experiments.

A matrix completion problem, which aims to recover a complete matrix from its partial observations, is one of the important problems in the machine learning field and has been studied actively. However, there is a discrepancy between the mainstream problem setting, which assumes continuous-valued observations, and some practical applications such as recommendation systems and SNS link predictions where observations take discrete or even binary values. To cope with this problem, Davenport et al. (2014) proposed a binary matrix completion (BMC) problem, where observations are quantized into binary values. Hsieh et al. (2015) proposed a PU (Positive and Unlabeled) matrix completion problem, which is an extension of the BMC problem. This problem targets the setting where we cannot observe negative values, such as SNS link predictions. In the construction of their method for this setting, they introduced a methodology of the classification problem, regarding each matrix entry as a sample. Their risk, which defines losses over unobserved entries as well, indicates the possibility of the use of unobserved entries. In this paper, motivated by a semi-supervised classification method recently proposed by Sakai et al. (2017), we develop a method for the BMC problem which can use all of positive, negative, and unobserved entries, by combining the risks of Davenport et al. (2014) and Hsieh et al. (2015). To the best of our knowledge, this is the first BMC method which exploits all kinds of matrix entries. We experimentally show that an appropriate mixture of risks improves the performance.