We consider the problem of approximate joint triangularization of a set of noisy jointly diagonalizable real matrices. Approximate joint triangularizers are commonly used in the estimation of the joint eigenstructure of a set of matrices, with applications in signal processing, linear algebra, and tensor decomposition. By assuming the input matrices to be perturbations of noise-free, simultaneously diagonalizable ground-truth matrices, the approximate joint triangularizers are expected to be perturbations of the exact joint triangularizers of the ground-truth matrices. We provide a priori and a posteriori perturbation bounds on the `distance' between an approximate joint triangularizer and its exact counterpart. The a priori bounds are theoretical inequalities that involve functions of the ground-truth matrices and noise matrices, whereas the a posteriori bounds are given in terms of observable quantities that can be computed from the input matrices. From a practical perspective, the problem of finding the best approximate joint triangularizer of a set of noisy matrices amounts to solving a nonconvex optimization problem. We show that, under a condition on the noise level of the input matrices, it is possible to find a good initial triangularizer such that the solution obtained by any local descent-type algorithm has certain global guarantees. Finally, we discuss the application of approximate joint matrix triangularization to canonical tensor decomposition and we derive novel estimation error bounds.

Early diagnosis, playing an important role in preventing progress and treating the Alzheimer's disease (AD), is based on classification of features extracted from brain images. The features have to accurately capture main AD-related variations of anatomical brain structures, such as, e.g., ventricles size, hippocampus shape, cortical thickness, and brain volume. This paper proposes to predict the AD with a deep 3D convolutional neural network (3D-CNN), which can learn generic features capturing AD biomarkers and adapt to different domain datasets. The 3D-CNN is built upon a 3D convolutional autoencoder, which is pre-trained to capture anatomical shape variations in structural brain MRI scans. Fully connected upper layers of the 3D-CNN are then fine-tuned for each task-specific AD classification. Experiments on the \emph{ADNI} MRI dataset with no skull-stripping preprocessing have shown our 3D-CNN outperforms several conventional classifiers by accuracy and robustness. Abilities of the 3D-CNN to generalize the features learnt and adapt to other domains have been validated on the \emph{CADDementia} dataset.

This letter presents the sparse vector signal detection from one bit compressed sensing measurements, in contrast to the previous works which deal with scalar signal detection. In this letter, available results are extended to the vector case and the GLRT detector and the optimal quantizer design are obtained. Also, a double-detector scheme is introduced in which a sensor level threshold detector is integrated into network level GLRT to improve the performance. The detection criteria of oracle and clairvoyant detectors are also derived. Simulation results show that with careful design of the threshold detector, the overall detection performance of double-detector scheme would be better than the sign-GLRT proposed in [1] and close to oracle and clairvoyant detectors. Also, the proposed detector is applied to spectrum sensing and the results are near the well known energy detector which uses the real valued data while the proposed detector only uses the sign of the data.

The Budapest BI Forum is the leading vendor-independent business intelligence and analytics conference in Hungary. One of the main feature of the event is the multi-subject approach: we have Pydata, Rstats, Data visualization, machine learning and BI talk in the 2 day s of the conference. This way speaking at the Budapest BI Forum in any of the tracks also provides a special chance to learn about other interesting fields of BI and analytics. This year we will have the following track, with a separate CFP for each. All speakers are welcome to submit more than one talks to the same or to different tracks.

Ben Dickson is a software engineer and freelance writer. He writes regularly on business, technology and politics. Thanks to technologies that generate, store and analyze huge sets of data, companies are able to perform tasks that previously were impossible. But the added benefit does come with its own setbacks, specifically from a security standpoint. With reams of data being generated and transferred over networks, cybersecurity experts will have a hard time monitoring everything that gets exchanged -- potential threats can easily go unnoticed.

It's your nighttime routine: You drop your phone onto the nightstand charging pad, and it asks about your day. You tell it, talking to the virtual personal assistant just like you'd talk to a friend. Your phone's artificial intelligence knows you almost as well as you know yourself (maybe even better). So when it suggests ways to get through tomorrow's calendar, you trust its advice. AI is practically everywhere, and getting smarter all the time.

Industrial robots used to be dumb, somewhat inflexible, and mostly blind - but also fast, precise and very efficient. As the cost of components, sensors and vision systems has been dropping, vision-enabled robots are becoming more prevalent and capable, and the industry is dramatically changing. Those changes can be seen in recent trends in China, investments in and acquisitions of robotic companies, by an analysis of recent startup companies, new and widening application areas for robot use, and technological developments. For the past 50 years industrial robots have picked the low-hanging fruit of manufacturing by handling the dull, dirty and dangerous tasks. But today, as consumers want more personalized products, and want them faster, and as costs have dropped and executives have pushed for greater productivity through automation, mobile and vision-enabled robots are emerging and being deployed in many new application areas, particularly for SMEs and in logistics, but also in government, agriculture, surveying, construction and healthcare.

After escalating one of the most lethal covert operations in U.S. history, President Obama finally made a public estimate of the civilian cost of the nation's secret drone program, which has targeted Islamic militants in remote corners of the globe. Human rights groups immediately challenged the estimate and the amount of transparency from the administration, saying both were too limited. The White House said that 64 to 116 civilians had been wrongly killed in 473 strikes launched by the U.S. government from the time Obama was inaugurated and the end of last year. The vast majority of the attacks were launched by drones, officials said, but the estimate also covers some strikes using manned aircraft. Monitoring organizations estimate the number of civilians killed in U.S. strikes ranges from 200 to more than 1,000.

BREAKING: After truck driver suspected Tesla driver in fatal crash was watching a video, police say DVD player found pic.twitter.com/qoZjd6jt62 When engaged, the Autopilot feature reminds drivers to keep their hands on the wheel and remain attentive, however many posted videos show that not everyone actually does that. From what we know about this particular crash, there are a number of elements at play. Slashgear points out that while Florida laws ban forbid TV screens visible to the driver while a car is in motion, the driver, identified as Joshua Brown, was from Ohio, which does not have a specific law against that. Mobileye, the company behind some of the sensor tech involved in the Tesla, said in a statement to Electrek that "today's collision avoidance technology, or Automatic Emergency Braking (AEB) is defined as rear-end collision avoidance, and is designed specifically for that. This incident involved a laterally crossing vehicle, which current-generation AEB systems are not designed to actuate upon."

BMW, a company that prides itself on building "the ultimate driving machine," plans to start producing fully autonomous vehicles by 2021 for ridesharing programs. Think of it as Uber for people who don't like people. This is a surprising move, given that the company has said essentially nothing about technology that everyone from Google to General Motors to Tesla is racing to develop. And it marks a radical departure from the slow-and-steady approach of the mainstream automakers, who see the technology rolling out slowly over the next two decades. Still, ze Germans see themselves surging ahead by relying upon help from Intel and Mobileye, an Israeli firm that dominates the market for the cameras that are key to active safety features like collision warning and lane-keeping.