In hopes of luring skilled workers both at home and form abroad with a chance to reside in Japan's ancient capital of Kyoto, Line Corp., a popular messaging app service, opened a new development base in the heart of the city in June. The firm plans to raise the number of engineers to 3,000, from around 2,100, in the near future. At present, the Kyoto office has 19 engineers of which 10 are from overseas. While more tech staffers are needed, "we won't lower our hiring bar," said Ryohei Miyota, who oversees the hiring of engineers at Line's Kyoto branch. Still, Line has shown it is willing to be flexible with other conditions.

MIYAZAKI – Like many of Japan's smaller cities, Miyazaki has been hit by a growing labor crunch, a trend highlighted by the mere 56.8 percent of high school graduates that chose to remain in the prefecture to work -- the third worst among the country's 47 prefectures. In the hard-hit information technology sector, the city has been encouraging firms to run businesses there to help energize the area, said Tsugunobu Ogino, president of KJS Co., a Miyazaki-based IT firm that makes e-learning systems. "But they are struggling to find engineers, since many move to Tokyo," he said. Now, the city in the southern Kyushu region may have found an unexpected solution, one thousands of kilometers away: Bangladesh. The South Asian nation faces a scenario that is almost the complete inverse of Japan -- there are simply not enough jobs for its ample working population.

ATAQ, YEMEN – The United States has waged a drone war in Yemen for 16 years, trying to suppress al-Qaida's branch there. But the campaign has had a hidden cost: civilians cut down by the drones' missiles. There is no comprehensive count of civilian deaths because of the difficulty of confirming identities and allegiances of those killed. But in an examination of drone strikes this year alone, AP found that at least 30 of the dead likely did not belong to al-Qaida. That is around a third of all those killed in drone strikes so far in 2018.

Officially, the group is responsible for issuing fatwas – which are rulings by Islamic authorities – in Iraqi Kurdistan. However, other imams have come out against the ban.

How will changing technology standards affect our ability to commercialize technology in different markets? "U.S. companies have long enjoyed the benefits of being able to define the global standards for technology – setting domestic standards for things from USB ports to more complex tech which would then be exported globally; this conferred great competitive advantage for companies here," the WCD notes. "But the power has shifted. China is investing billions as a nation, coordinating and creating the standards for all tech and AI by 2030. In local industries and at the enterprise level, they are setting the stage for what it means to work and be competitive, and also setting the standards within critical infrastructure, energy, utilities, and transportation. All of this affects how companies in the U.S. will be able to take their products to market overseas."

Model-free reinforcement learning methods such as the Proximal Policy Optimization algorithm (PPO) have successfully applied in complex decision-making problems such as Atari games. However, these methods suffer from high variances and high sample complexity. On the other hand, model-based reinforcement learning methods that learn the transition dynamics are more sample efficient, but they often suffer from the bias of the transition estimation. How to make use of both model-based and model-free learning is a central problem in reinforcement learning. In this paper, we present a new technique to address the trade-off between exploration and exploitation, which regards the difference between model-free and model-based estimations as a measure of exploration value. We apply this new technique to the PPO algorithm and arrive at a new policy optimization method, named Policy Optimization with Model-based Explorations (POME). POME uses two components to predict the actions' target values: a model-free one estimated by Monte-Carlo sampling and a model-based one which learns a transition model and predicts the value of the next state. POME adds the error of these two target estimations as the additional exploration value for each state-action pair, i.e, encourages the algorithm to explore the states with larger target errors which are hard to estimate. We compare POME with PPO on Atari 2600 games, and it shows that POME outperforms PPO on 33 games out of 49 games.

The past few years have witnessed the fast development of different regularization methods for deep learning models such as fully-connected deep neural networks (DNNs) and Convolutional Neural Networks (CNNs). Most of previous methods mainly consider to drop features from input data and hidden layers, such as Dropout, Cutout and DropBlocks. DropConnect select to drop connections between fully-connected layers. By randomly discard some features or connections, the above mentioned methods control the overfitting problem and improve the performance of neural networks. In this paper, we proposed two novel regularization methods, namely DropFilter and DropFilter-PLUS, for the learning of CNNs. Different from the previous methods, DropFilter and DropFilter-PLUS selects to modify the convolution filters. For DropFilter-PLUS, we find a suitable way to accelerate the learning process based on theoretical analysis. Experimental results on MNIST show that using DropFilter and DropFilter-PLUS may improve performance on image classification tasks.

An efficient state estimation model, neural network estimation (NNE), empowered by machine learning techniques, is presented for full quantum state tomography (FQST). A parameterized function based on neural network is applied to map the measurement outcomes to the estimated quantum states. Parameters are updated with supervised learning procedures. From the computational complexity perspective our algorithm is the most efficient one among existing state estimation algorithms for full quantum state tomography. We perform numerical tests to prove both the accuracy and scalability of our model.

Last week, we featured sales on gear from outdoor retailers who were hoping to get a jump on the creeping monolith that is Black Friday. Many other manufacturers and retailers are hoping to entice you into checking out their wares before you fall into a post-turkey stupor. If you're in the mood--or if you have better things to do on your holiday weekend than shop--we've rounded up more than a dozen of our favorite deals to check out. It may not be as beautiful as the Nest, but the Ecobee4 was our smart home thermostat of choice when we reviewed it last year. Temperature sensors ensure that the far reaches of your house remain unbaked (and unfrozen).

OSAKA – Despite advances in facial recognition technology, the police in Osaka still rely on pure skill to find fugitives, with investigators using only their memory to arrest dozens of wanted criminals every year. While other police forces in the world have "super recognizer" units that hunt down fugitives, the so-called miatari (look and hit) technique used in Osaka has contributed to the arrests of over 4,000 criminals in Japan since the Osaka Prefectural Police introduced it as a formal investigative method in November 1978. There has not been a single wrongful arrest. "The best part of this method is being able to detect fugitives who are hard to find in normal investigations," said a senior investigator in Osaka. He says a forensic analysis is an imperative part of criminal investigations, but "we want to pass on the tradition because our job is to make sure no one gets away with a crime."