Task-oriented dialog systems are becoming pervasive, and many companies heavily rely on them to complement human agents for customer service in call centers. With globalization, the need for providing cross-lingual customer support becomes more urgent than ever. However, cross-lingual support poses great challenges---it requires a large amount of additional annotated data from native speakers. In order to bypass the expensive human annotation and achieve the first step towards the ultimate goal of building a universal dialog system, we set out to build a cross-lingual state tracking framework. Specifically, we assume that there exists a source language with dialog belief tracking annotations while the target languages have no annotated dialog data of any form. Then, we pre-train a state tracker for the source language as a teacher, which is able to exploit easy-to-access parallel data. We then distill and transfer its own knowledge to the student state tracker in target languages. We specifically discuss two types of common parallel resources: bilingual corpus and bilingual dictionary, and design different transfer learning strategies accordingly. Experimentally, we successfully use English state tracker as the teacher to transfer its knowledge to both Italian and German trackers and achieve promising results.

A new facial recognition technology caught a man trying to enter the US using a passport belonging to someone else, US officials say. Officials with the US Customs and Border Protection (CBP) and the Office of Field Operations (OFO) intercepted a 26-year-old man, the agencies referred to as an "imposter", who reportedly attempted to use a French passport belonging to someone else, at Washington's Dulles International Airport. The man was travelling to the US from Brazil. "The officer utilised CBP's new facial comparison biometric technology which confirmed the man was not a match to the passport he presented," the CBP press release read. It added: "A search revealed the man's authentic Republic of Congo identification card concealed in his shoe."

LG's deep collaboration with Google continues, as it just announced that Google Assistant is coming to ThinQ smart TVs in seven more markets and five languages. LG will also expand Amazon Alexa support to Australia and Canada. Google Assistant first arrived to ThinQ AI TVs in the US earlier this year. Google Assistant on LG ThinQ TVs lets you control smart lights, appliances, robotic vacuums and other home devices, check the weather, make a restaurant reservation and more. You can also shout at the TV (via the "mic" button on the remote) or Google Home devices to pause ThinQ TV programming, change channels, lower the volume, etc. Google Home support for ThinQ TVs works now in Canada, Australia and the UK, and non-English language support for France, Spain, South Korea and Germany will arrive by the end of 2018.

Russian company Kalashnikov have released a robot - and it's not quite what anyone expected. 'Igorek' ('little Igor', not its official name) was revealed to the public on Monday in Moscow by the company behind the famous AK-47 gun which has sold more than 100 million units worldwide. The 13-feet (3.96m) tall, 4.5-tonne, manned robot is designed for "carrying out engineering and combat tasks", according to Kalashnikov, at the ARMY Forum. The Forum describes itself as the "world's leading exhibition of arms and military equipment, the authoritative platform for discussing innovative ideas and developments for the armed forces". For the moment, however, Igorek is completely immobile.

Abstract-- Extremely preterm infants often require endotracheal intubation and mechanical ventilation during the first days of life. Due to the detrimental effects of prolonged invasive mechanical ventilation (IMV), clinicians aim to extubate infants as soon as they deem them ready. Unfortunately, existing strategies for prediction of extubation readiness vary across clinicians and institutions, and lead to high reintubation rates. We present an approach using Random Forest classifiers for the analysis of cardiorespiratory variability to predict extubation readiness. We address the issue of data imbalance by employing random undersampling of examples from the majority class before training each Decision Tree in a bag. By incorporating clinical domain knowledge, we further demonstrate that our classifier could have identified 71% of infants who failed extubation, while maintaining a success detection rate of 78%.

After birth, extremely preterm infants often require specialized respiratory management in the form of invasive mechanical ventilation (IMV). Protracted IMV is associated with detrimental outcomes and morbidities. Premature extubation, on the other hand, would necessitate reintubation which is risky, technically challenging and could further lead to lung injury or disease. We present an approach to modeling respiratory patterns of infants who succeeded extubation and those who required reintubation which relies on Markov models. We compare the use of traditional Markov chains to semi-Markov models which emphasize cross-pattern transitions and timing information, and to multi-chain Markov models which can concisely represent non-stationarity in respiratory behavior over time. The models we developed expose specific, unique similarities as well as vital differences between the two populations.

The news has always played an essential role in our lives, keeping us informed about the world and the issues we care about. But the fact is, while there is a sea bad news, there is also a plethora of "good news" happening where people are making progress solving real issues. The Google Assistant is now making this kind of news easier to find. "Tell me something good" is a new experimental feature for Assistant users in the US that delivers your daily dose of good news. Just say "Hey Google, tell me something good" to receive a brief news summary about people who are solving problems for our communities and our world, the company said in a blog post.

Driverless cars could make our roads safer and reduce congestion. But the algorithms driving them will also have to make life-or-death decisions. At some stage in the future, a fully autonomous car may determine who lives and who dies on our roads. These machines are being tested right now and Australian politicians are looking overseas for leadership, emboldened by the promise of fewer fatalities and less congestion. At the moment, there must be a human behind the wheel of these cars at all times, but government agencies are already working on a legal framework for when machines are totally in control.

Curtin University, in Western Australia, will be working with Optus Business as they form a research group that will focus on the impact of artificial intelligence (AI) on regional telecommunications, higher education and the urban environment. According to the report made by the University, an artificial intelligence research group will be formed from the five-year alliance. The group will be embedded in the School of Electrical Engineering, Computing and Mathematical Sciences at the University, having strong links to the Curtin Institute for Computation. The excellent research, teaching and learning capabilities of the University will be synergised with the market-leading technology and infrastructure capabilities of the telco company and will be fully leveraged by the alliance of both. The research group will involve the appointment of an Optus Chair in Artificial Intelligence and three Optus Research Fellows focusing on applying artificial intelligence technologies in areas such as regional telecommunications, improving higher education student outcomes and the urban environment.

Localization of chest pathologies in chest X-ray images is a challenging task because of their varying sizes and appearances. We propose a novel weakly supervised method to localize chest pathologies using class aware deep multiscale feature learning. Our method leverages intermediate feature maps from CNN layers at different stages of a deep network during the training of a classification model using image level annotations of pathologies. During the training phase, a set of \emph{layer relevance weights} are learned for each pathology class and the CNN is optimized to perform pathology classification by convex combination of feature maps from both shallow and deep layers using the learned weights. During the test phase, to localize the predicted pathology, the multiscale attention map is obtained by convex combination of class activation maps from each stage using the \emph{layer relevance weights} learned during the training phase. We have validated our method using 112000 X-ray images and compared with the state-of-the-art localization methods. We experimentally demonstrate that the proposed weakly supervised method can improve the localization performance of small pathologies such as nodule and mass while giving comparable performance for bigger pathologies e.g., Cardiomegaly