Buried landmines and unexploded remnants of war are a constant threat for the population of many countries that have been hit by wars in the past years. The huge amount of human lives lost due to this phenomenon has been a strong motivation for the research community toward the development of safe and robust techniques designed for landmine clearance. Nonetheless, being able to detect and localize buried landmines with high precision in an automatic fashion is still considered a challenging task due to the many different boundary conditions that characterize this problem (e.g., several kinds of objects to detect, different soils and meteorological conditions, etc.). In this paper, we propose a novel technique for buried object detection tailored to unexploded landmine discovery. The proposed solution exploits a specific kind of convolutional neural network (CNN) known as autoencoder to analyze volumetric data acquired with ground penetrating radar (GPR) using different polarizations. This method works in an anomaly detection framework, indeed we only train the autoencoder on GPR data acquired on landmine-free areas. The system then recognizes landmines as objects that are dissimilar to the soil used during the training step. Experiments conducted on real data show that the proposed technique requires little training and no ad-hoc data pre-processing to achieve accuracy higher than 93% on challenging datasets.

Modern data analysis pipelines are becoming increasingly complex due to the presence of multi-view information sources. While graphs are effective in modeling complex relationships, in many scenarios a single graph is rarely sufficient to succinctly represent all interactions, and hence multi-layered graphs have become popular. Though this leads to richer representations, extending solutions from the single-graph case is not straightforward. Consequently, there is a strong need for novel solutions to solve classical problems, such as node classification, in the multi-layered case. In this paper, we consider the problem of semi-supervised learning with multi-layered graphs. Though deep network embeddings, e.g. DeepWalk, are widely adopted for community discovery, we argue that feature learning with random node attributes, using graph neural networks, can be more effective. To this end, we propose to use attention models for effective feature learning, and develop two novel architectures, GrAMME-SG and GrAMME-Fusion, that exploit the inter-layer dependencies for building multi-layered graph embeddings. Using empirical studies on several benchmark datasets, we evaluate the proposed approaches and demonstrate significant performance improvements in comparison to state-of-the-art network embedding strategies. The results also show that using simple random features is an effective choice, even in cases where explicit node attributes are not available.

We present a method for Temporal Difference (TD) learning that addresses several challenges faced by robots learning to navigate in a marine environment. For improved data efficiency, our method reduces TD updates to Gaussian Process regression. To make predictions amenable to online settings, we introduce a sparse approximation with improved quality over current rejection-based sparse methods. We derive the predictive value function posterior and use the moments to obtain a new algorithm for model-free policy evaluation, SPGP-SARSA. With simple changes, we show SPGP-SARSA can be reduced to a model-based equivalent, SPGP-TD. We perform comprehensive simulation studies and also conduct physical learning trials with an underwater robot. Our results show SPGP-SARSA can outperform the state-of-the-art sparse method, replicate the prediction quality of its exact counterpart, and be applied to solve underwater navigation tasks.

The FBI has ordered a suspect to unlock his iPhone X using the facial recognition feature in the first case known worldwide of authorities using Apple's face ID technology to pry into devices. The incident occurred in Columbus, Ohio, when the FBI entered the home of Grant Michalski, 28, on August 10 while investigating him for child abuse. An agent told Michalski's to put his face to the phone and once inside uncovered salacious chats that helped charge him with receiving and possessing child pornography, according toForbes. Forbes obtained court documents which revealed special agent David Knight entered Michalski's home with a search warrant and forced the suspect to unlock his phone with his face. This allowed Knight to go through the contents of Michalski's device, including online chats and photos.

In the U.S. for-profit, insurance-driven healthcare system a conundrum arises, which presents a complex challenge, with health plans enroll beneficiaries who are covered by both Medicare and Medicaid. These people are known as dually eligible beneficiaries. The proportion of people who are Medicare beneficiaries and who are also enrolled in Medicaid varies at any given time, but it runs into several million people. Medicare is a national health insurance program, administered by the Centers for Medicaid and Medicare Services of the U.S. federal government. It provides health insurance for Americans aged 65 and older who have worked and paid into the system through the payroll tax.

Good judgement and skills such as emotional intelligence, empathy, creativity and wisdom will become even more important to lawyering as the use of artificial intelligence increases, says UNSW Professor of Law Michael Legg. The director of the Law Society of NSW Future of Law and Innovation in the Profession (FLIP) stream at UNSW Law discussed what an AI-enhanced lawyer looks like at the Law Society's FLIP Conference in Sydney this month. "Whatever the nature of their practice, it has been said that the most important skill of lawyering is sound judgement," Professor Legg said. "Sound judgement is about more than answering legal questions – it encapsulates the relational and contextual elements of being a'problem-solving' lawyer. "None of the AI technologies currently available have the capacity to completely replace lawyers, as each still requires the exercise of human judgement as part of the process." The talk was part of a collaboration on research between UNSW Law and the NSW Law Society, which is responding to issues such as legal technology and new ways of working raised by the NSW Law Society's FLIP report in 2016. The report recognised that the legal profession is undergoing change at a pace never before experienced and in unforeseen ways, which has major ramifications for the legal profession, clients and society, particularly in relation to access to justice. FLIP Stream Research Fellow Dr Felicity Bell showed research by Israel-based contract review platform LawGeex, which compared the efforts of 20 "experienced" lawyers in reviewing five different non-disclosure agreements, with those of its software. The lawyers took, on average, 92 minutes to review all five contracts, while the software took a mere 26 seconds. "The software outperformed the lawyers in terms of accuracy, and vastly outperformed them in terms of speed," she said. "The lawyers took, on average, 92 minutes to review all five contracts, while the software took a mere 26 seconds.

Republican Arizona Sen. Jeff Flake has developed a reputation as someone who is willing to criticize Donald Trump and his party enablers in unusually blunt terms--but not generally willing to use his status as a potential swing vote in a narrowly divided Senate to initiate investigations into administration corruption, protect Robert Mueller's special counsel investigation, or achieve any of the other good-governance goals he might be expected to support given his feelings on Trumpism. That was Flake's reputation, at least, before he announced a dramatic effort last Friday to delay Brett Kavanaugh's confirmation vote until the FBI could conduct a weeklong review of sexual assault accusations against the Supreme Court nominee. When it became clear that fellow Trump-skeptic GOP senators Susan Collins and Lisa Murkowski supported Flake's position, the White House almost immediately agreed to order just such an investigation. Over the weekend, though, there was some grinding in the gears. NBC reported, and other outlets confirmed, that the White House had set narrow limits on which individuals the FBI was allowed to contact.

You lock your phone so other people can't access it. But how you lock your phone is an important factor in whether law enforcement can compel you to unlock it. Apple's year-old Face ID system is no exception. On Sunday, Forbes reported the first known example of law enforcement anywhere using a suspect's face to unlock a phone during an investigation. The question of whether cops can force someone to unlock their phone in the US for a search hinges on Fifth Amendment protects against self-incrimination--that no one "shall be compelled in any criminal case to be a witness against" themselves. Privacy advocates argue that this extends to the act of unlocking a phone, or generally decrypting data on a device.

I was invited to give a keynote lecture at the XV International Conference on University Libraries in Mexico City last month. The conference debated the changes occurring or need to occur in university libraries towards the United Nations' "Agenda 2030" horizon. My lecture, titled "Libraries and Knowledge in the Age of Information Overload", took a close look at the impact of the rampant Info Overload that is affecting the entire knowledge landscape on the academic library, and vice versa. The main thing is that where in centuries past knowledge – carried in books – was scarce and precious, today the internet has created a huge overload of information, dubious information and fake information; we are all struggling to survive this info tsunami. As a result, if the libraries of the past were like serene shrines where people would come to bask in the glow of accumulated knowledge, today libraries must serve as sanctuaries where users can get help to survive the aforementioned tsunami, and gain the skills and support to allow them to find insight yet discriminate against the fake and the useless.