We tackle unsupervised anomaly detection (UAD), a problem of detecting data that significantly differ from normal data. UAD is typically solved by using density estimation. Recently, deep neural network (DNN)-based density estimators, such as Normalizing Flows, have been attracting attention. However, one of their drawbacks is the difficulty in adapting them to the change in the normal data's distribution. To address this difficulty, we propose AdaFlow, a new DNN-based density estimator that can be easily adapted to the change of the distribution. AdaFlow is a unified model of a Normalizing Flow and Adaptive Batch-Normalizations, a module that enables DNNs to adapt to new distributions. AdaFlow can be adapted to a new distribution by just conducting forward propagation once per sample; hence, it can be used on devices that have limited computational resources. We have confirmed the effectiveness of the proposed model through an anomaly detection in a sound task. We also propose a method of applying AdaFlow to the unpaired cross-domain translation problem, in which one has to train a cross-domain translation model with only unpaired samples. We have confirmed that our model can be used for the cross-domain translation problem through experiments on image datasets.

This technology is anything but new, but it's never gotten as much attention as it does today. In recent years, AI has not only shown real technological progress but also shows how it can be used for political purposes. While the current AI race cannot really be compared to the nuclear race of yesteryear, we can nevertheless see that the major powers are seeking to control it in order to have an economic, social and military advantage over other countries. The USA and China are at the forefront of this race, of course, and the EU is well aware of this. It has therefore decided to react to better equip itself.

Today, Google shared information about some of the AI work it's doing in Asia, but in a blog post about the work, it also made a pretty clear statement about how its facial recognition technology will and won't be used for the time being. The company noted that while facial recognition systems stand to be quite useful in a variety of situations, from assistive technologies to locating missing people, they also comes with risks. "Like many technologies with multiple uses, facial recognition merits careful consideration to ensure its use is aligned with our principles and values, and avoids abuse and harmful outcomes," Google said. "We continue to work with many organizations to identify and address these challenges, and unlike some other companies, Google Cloud has chosen not to offer general-purpose facial recognition APIs before working through important technology and policy questions." Facial recognition technology has come under the spotlight in recent years, with everyone from local law enforcement to Taylor Swift employing it in some way.

In science news around the world, China's largest research funding agency expresses support for the goals of Plan S, the push by European science funders for immediate open access to research publications. For the second year in a row, carbon dioxide emissions from fossil fuels are projected to hit a new high, growing 2.7% this year. NASA's Voyager 2 probe becomes only the second humanmade object to enter interstellar space. Empathy expert Tania Singer resigns as director of the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig, Germany, after a commission confirmed allegations of bullying. Dozens of African researchers are denied visas for an artificial intelligence (AI) meeting in Montreal, Canada, even as the Canadian government takes steps to advance the country's standing in AI.

Technology researchers in China have been ordered to not travel to the US unless it is absolutely necessary, amid rising tensions between the two countries. Staff working in sensitive tech sectors were given the warning following the arrest of a Chinese tech executive in Canada, a source told the South China Morning Post. Workers at a research agency were also told to remove any sensitive data from laptops, mobile phones and other devices if travel to the US was essential. The warning comes after a similar order from US tech giant Cisco to some of its employees, which asked them the to any non-essential travel to China. Cisco has since said the email to its employees was "sent in error."

One of the stand-out features of Samsung's forthcoming flagship smartphone, presumably called the Galaxy S10, is expected to be the camera, with the latest rumour suggesting an ultra-wide lens that will significantly improve image quality. The new camera feature was spotted in a beta version of the Android Pie mobile operating system, which includes an option for "Ultra wide lens correction" within the camera settings. Several rumours surrounding the Galaxy S10 camera have already hinted at a multi-lens setup, both on the front and rear of the device. Last month, a report claimed that the successor to the Galaxy S9 will feature twice as many cameras as Apple's latest iPhone, with two on the front and four on the rear. The Galaxy S10, set to launch on the 10th anniversary of Samsung's first Galaxy flagship smartphone, is also expected to shun the precedent set by Apple by using a "hole-punch" instead of notch in order to incorporate a front-facing camera.

It's the feature most players will have been waiting for so long they'll never actually get to use it, but Pokemon Go has finally allowed players to battle in the game. The introduction of player versus player fights arrives in the latest update to the game and lets Pokemon trainers use the creatures and skills they've honed over more than two years to take on friends. Battles had been a notably missing part of the game since it was released in the summer of 2016. Now developers say they have finally arrived, with players receiving special rewards for playing them. Developers are clear that the matches won't only reward players who have the strongest Pokemon and the best moves, privileging those who have been playing for a while.

The surprise arrest of Huawei Technologies Co.'s Chief Financial Officer Meng Wanzhou is about to impact one of the Chinese company's suppliers in Japan. Yaskawa Electric Corp., which supplies industrial robots for Huawei's smartphone and telecom gear factories, saw all orders for its machines put on hold after the arrest, President Hiroshi Ogasawara said in an interview Wednesday. Of Yaskawa's ¥448.5 billion ($4 billion) in revenue for the fiscal year that ended in February, 23 percent came from China. "My people on the ground in China say that Huawei is turned upside down internally," Ogasawara said. "All kinds of capex deals are temporarily on hold as they figure things out."

A team of machine learning developers has created a system for making caricatures of people's faces. In a caricature, the artist creates a drawing of a face, with different parts of it greatly enlarged, or exaggerated, in other ways. The pictures or images are usually made to make the person look funny. Such drawings can be difficult for machines to produce. This is often because the human face is made up of complex shapes with a lot of extremely small details.

It's not every day that TIME magazine calls you a genius. Walach is the 30-year-old CEO of Aidoc, a two-year-old Tel Aviv-based startup that is saving lives through medical imaging. Aidoc applies proprietary artificial intelligence to the millions of images generated every year by CT scans in order to catch serious issues before a human radiologist even has a chance to review the results. Aidoc has already received US and European approval to assess scans of brain hemorrhages and spinal fractures. TIME included the startup on its list of "50 Genius Companies of 2018," a prestigious cohort that includes well-known names such as Amazon, Airbnb and Apple (and that's just the As).