"Through an AI theoretical analysis of past data, we will be able to properly respond to cases without just relying on teachers' past experiences," Otsu Mayor Naomi Koshi said of the planned analysis, set to begin from the next fiscal year. AI will be used to analyze 9,000 suspected bullying cases reported by elementary and junior high schools in the city over the six years through fiscal 2018. It will examine the school grade and gender of the suspected victims and perpetrators as well as when and where the incidents occurred. Statistical analysis of the data is expected to help local authorities and teachers identify forms of bullying that tend to escalate in seriousness and which therefore require extra attention, the Otsu board of education said. The AI analysis will also look at other factors, such as school absenteeism and academic achievement, and the findings will be compiled into a report for use by teachers and in training seminars.

Apple has finally released a software update that fixes maybe its creepiest issue ever: a FaceTime bug that allowed people to snoop on others through their phone. Downloading the new version of iOS – named iOS 12.1.4 The new version includes a message that it "provides important security updates and is recommended for all users", and doesn't make any mention of new features or any other changes. Apple has also turned back on the servers that power the group FaceTime feature, which were unplugged as it attempted to stop people exploiting the bug. For now, phones and other devices such as Macs that have not been updated will not be able to use the feature until they download the new software.

In March 2011, one of the most powerful earthquakes recorded in Japan shook the country's northeastern coast. It flattened thousands of homes and buildings, destroyed infrastructure, and triggered a tsunami that wiped out whole communities. As the death toll rose above 20,000 with many more people injured and homeless, scientists and policymakers around the globe asked: How can society use technology to better prepare for, and respond to, natural disasters? How can we work faster to save lives? Leading this new thinking are researcher Yanbing Bai and a team at Tohoku University's International Research Institute of Disaster Science based in the region hardest hit by the quake.

As the smartphone falls in price while its capabilities improve, it is becoming a valuable tool in the diagnosis of a growing number of diseases and ailments around the world. When Yonatan Adiri's mother fell down a bank and briefly lost consciousness when travelling in China, an initial diagnosis suggested she had a few broken ribs, but nothing more serious. Doctors were keen to fly her to Hong Kong for treatment. But Yonatan's father was worried and took photos of the CT [computerised tomography] scans of the injuries, emailing them to his son. Yonatan showed the images to a trauma doctor, who instantly diagnosed a punctured lung.

It's common to say that China and the US are in an "AI arms race," as each country races to develop and commercialize deep learning and other AI technologies before the other. A new report shows that a more literal AI arms race is also under way. View from Beijing: The new report, from the Center for a New American Security, an influential think tank, offers an unusual level of access. It draws on various conferences as well as meetings with officials in China's Ministry of Foreign Affairs. Military mind-set: The report says that China is moving quickly to add ever more AI and autonomy to military weapons systems.

WASHINGTON - A small tropical reef fish was able to recognize itself in a mirror, scientists said Thursday, in a finding that raises provocative questions about assessing self-awareness and cognitive abilities in animals. The study involved experiments in which the bluestreak cleaner wrasse was given a mirror self-recognition test, a technique developed in 1970 for gauging animal self-awareness. In aquarium experiments at Osaka City University, researchers applied a brown-colored mark on the fish's body in a place that could be seen only in a mirror reflection. The fish tried to remove the marks by scraping their bodies on hard surfaces after watching themselves in a mirror, but never tried to remove them without a mirror present, indicating they understood the reflection was of them, the researchers said. When a transparent, rather than brown, mark was applied, the fish never tried to remove it.

The European Commission's Joint Research Center released a report that explores the European perspective on artificial intelligence (AI), along with the global AI landscape's state of play. The report recognizes the value of AI across industry, but while acknowledging the fierce competition on AI taking place between the EU, China, and the United States, it ultimately dismisses the need for Europe to win this global race, arguing instead that for the EU, the more important goal is focusing on developing values and ethics in AI. This is a naive perspective, especially given that China is not only fiercely competing on developing AI, but also aspiring to dominate in AI so as to compete in industries where Europe is leading today. Ironically, even if the EU's first priority is to shape the values and ethics of AI, it will be severely limited in its ability to do so if it is not leading the development and adoption of this technology. Europe would be wrong to forget that any competition involves winner and losers--and more often than not, the winners are those who compete with gusto.

This paper presents a new algorithm, Evolutionary eXploration of Augmenting Memory Models (EXAMM), which is capable of evolving recurrent neural networks (RNNs) using a wide variety of memory structures, such as Delta-RNN, GRU, LSTM, MGU and UGRNN cells. EXAMM evolved RNNs to perform prediction of large-scale, real world time series data from the aviation and power industries. These data sets consist of very long time series (thousands of readings), each with a large number of potentially correlated and dependent parameters. Four different parameters were selected for prediction and EXAMM runs were performed using each memory cell type alone, each cell type with feed forward nodes, and with all possible memory cell types. Evolved RNN performance was measured using repeated k-fold cross validation, resulting in 1210 EXAMM runs which evolved 2,420,000 RNNs in 12,100 CPU hours on a high performance computing cluster. Generalization of the evolved RNNs was examined statistically, providing interesting findings that can help refine the RNN memory cell design as well as inform future neuro-evolution algorithms development.

Tensor factorization has become an increasingly popular approach to knowledge graph completion(KGC), which is the task of automatically predicting missing facts in a knowledge graph. However, even with a simple model like CANDECOMP/PARAFAC(CP) tensor decomposition, KGC on existing knowledge graphs is impractical in resource-limited environments, as a large amount of memory is required to store parameters represented as 32-bit or 64-bit floating point numbers. This limitation is expected to become more stringent as existing knowledge graphs, which are already huge, keep steadily growing in scale. To reduce the memory requirement, we present a method for binarizing the parameters of the CP tensor decomposition by introducing a quantization function to the optimization problem. This method replaces floating point-valued parameters with binary ones after training, which drastically reduces the model size at run time. We investigate the trade-off between the quality and size of tensor factorization models for several KGC benchmark datasets. In our experiments, the proposed method successfully reduced the model size by more than an order of magnitude while maintaining the task performance. Moreover, a fast score computation technique can be developed with bitwise operations.

Link prediction requires predicting which new links are likely to appear in a graph. Being able to predict unseen links with good accuracy has important applications in several domains such as social media, security, transportation, and recommendation systems. A common approach is to use features based on the common neighbors of an unconnected pair of nodes to predict whether the pair will form a link in the future. In this paper, we present an approach for link prediction that relies on higher-order analysis of the graph topology, well beyond common neighbors. We treat the link prediction problem as a supervised classification problem, and we propose a set of features that depend on the patterns or motifs that a pair of nodes occurs in. By using motifs of sizes 3, 4, and 5, our approach captures a high level of detail about the graph topology within the neighborhood of the pair of nodes, which leads to a higher classification accuracy. In addition to proposing the use of motif-based features, we also propose two optimizations related to constructing the classification dataset from the graph. First, to ensure that positive and negative examples are treated equally when extracting features, we propose adding the negative examples to the graph as an alternative to the common approach of removing the positive ones. Second, we show that it is important to control for the shortest-path distance when sampling pairs of nodes to form negative examples, since the difficulty of prediction varies with the shortest-path distance. We experimentally demonstrate that using off-the-shelf classifiers with a well constructed classification dataset results in up to 10 percentage points increase in accuracy over prior topology-based and feature learning methods.