A network of independently trained Gaussian processes (StackedGP) is introduced to obtain predictions of quantities of interest with quantified uncertainties. The main applications of the StackedGP framework are to integrate different datasets through model composition, enhance predictions of quantities of interest through a cascade of intermediate predictions, and to propagate uncertainties through emulated dynamical systems driven by uncertain forcing variables. By using analytical first and second-order moments of a Gaussian process with uncertain inputs using squared exponential and polynomial kernels, approximated expectations of quantities of interests that require an arbitrary composition of functions can be obtained. The StackedGP model is extended to any number of layers and nodes per layer, and it provides flexibility in kernel selection for the input nodes. The proposed nonparametric stacked model is validated using synthetic datasets, and its performance in model composition and cascading predictions is measured in two applications using real data.

Artificial intelligence (AI) and machine learning for security may be soon reaching a tipping point: In a global survey of C-suite executives, Radware found that about four in five (81%) reported having already or recently implemented more reliance on automated solutions. Some 57% of executives report trusting automated systems that employ AI and machine learning as much or more than humans to protect their organizations. Two in five (38%) executives indicated that within two years, automated security systems would be the primary resource for managing cybersecurity. "Businesses have to fight fire with fire," said Carl Herberger, vice president of security solutions at Radware. "Today's threat actors continue to build highly automated and adaptive tools, like the Mirai and Hajime botnets. These attacks can wreak catastrophic damage to a network. Executives that aren't yet fighting these new dynamic threats with continuously adaptive attack detection and mitigation capabilities are putting their organization at risk."

It's often been observed that technological advancement is a double-edged sword – and this is especially true in the realm of cybersecurity. Attackers have been known to use automation technology to stage and sustain their strikes – while those same machine learning algorithms and real-time response mechanisms can help enterprises that suffer an assault to speed up their efforts at remediation. However, this can only happen if these automated and learning technologies are actually being deployed by enterprise users. With recent security research suggesting that a typical business organization takes an average of 146 days to fix critical software or system vulnerabilities, the penetration of such technology clearly hasn't reached acceptable levels, yet. Nevertheless, there's an increasing level of buzz over machine learning, adaptive technologies, and Artificial Intelligence (AI) – and the roles they can and should play in improving enterprise cybersecurity.

Gov. Greg Abbott signed a bill Thursday that signals to Google, Uber and carmakers that they are welcome to test self-driving cars on the state's roads and highways without a driver behind the wheel. There was nothing in existing law that banned autonomous vehicles from Texas roads. After all, Google has been testing them since 2015 in Austin, and Arlington is rolling them out. And several Texas sites were chosen by the U.S. Department of Transportation to test the technology in closed-course settings. Yet because state statutes didn't address the emerging technology at all, some manufacturers have told state officials they were wary about testing vehicles alongside street and highway traffic in Texas.

With Yahoo now part of the Verizon empire, Jefferson Graham takes a good look at the homepage, and finds it....rather dated. This week Yahoo, one of the oldest Internet brands, became part of the Verizon empire, one of the companies (along with HuffPost, AOL and TechCrunch) that live within the newly created Oath unit, which hopes to compete against Google and Facebook as a bulked-up alternative for online advertisers. This Sept. 23, 2016 file photo shows the Yahoo logo pictured on a computer monitor in Taipei, Taiwan. Verizon on Tuesday, June 13, closed its $4.48 billion acquisition of Yahoo. And let's take a quick minute and look at the jewel of Oath, the beleaguered and ignored Yahoo.com.

For much of the past year, Uber has been caught up in increasing waves of scandal. Earlier this week, the backlash against Uber finally came to a head for the ride-hailing company. CEO Travis Kalanick announced Tuesday he plans to take an indefinite leave of absence from the company, saying in an internal email he needed "to work on myself" and "focus on building out a world-class leadership team." Last month, Kalanick's mother died in a boating accident that also sent his father to the hospital. For the moment, Kalanick leaves behind a company that is reeling from a string of public embarrassments and now is without its central driver.

Jobs are disappearing, incomes retreating, the precariat growing. Thousands of people risk their lives in stormy seas to flee wars, moribund economies and climate change on a daily basis. Traditional politicians continue to avoid publicly addressing the tsunami of unemployment, apparently baffled as to how to react to a historic transition: the automation of critical masses of labour once performed by humans. Five acronyms - AI, AR, VR, BC and UBI - promise to shape the developed world's future and solve the problems of the present. In the process, however, these innovations risk transforming the world around us, and upsetting humanity's very definition of itself.

Yemen-based AQAP has exploited a more than two-year-old civil war between the Iran-aligned Houthis and President Abd Rabbu Mansour Hadi's Saudi-backed government to enhance its influence in the impoverished country.

This week's news reminds me how much fun it is to be surprised by technology. Yesterday, the Paris-based AI startup raised $13 million, on top of an earlier $8 million investment, for technology that let developers put a voice assistant on nearly any device. Add this to recent advances from Amazon Alexa, Google Assistant, and Apple Siri, and it's obvious that voice is becoming the new interface much sooner than many people, including yours truly, ever anticipated. These are exponential leaps forward in the steady progress from command-based interfaces to conversational ones. It's as if the machines themselves are disappearing -- the "thing" we're conversing with is some crazy fantastic blend of artificial intelligence, super computer, bandwidth, and what have you, that we never see.

Spiking neural networks (SNNs) could play a key role in unsupervised machine learning applications, by virtue of strengths related to learning from the fine temporal structure of event-based signals. However, some spike-timing-related strengths of SNNs are hindered by the sensitivity of spike-timing-dependent plasticity (STDP) rules to input spike rates, as fine temporal correlations may be obstructed by coarser correlations between firing rates. In this article, we propose a spike-timing-dependent learning rule that allows a neuron to learn from the temporally-coded information despite the presence of rate codes. Our long-term plasticity rule makes use of short-term synaptic fatigue dynamics. We show analytically that, in contrast to conventional STDP rules, our fatiguing STDP (FSTDP) helps learn the temporal code, and we derive the necessary conditions to optimize the learning process. We showcase the effectiveness of FSTDP in learning spike-timing correlations among processes of different rates in synthetic data. Finally, we use FSTDP to detect correlations in real-world weather data from the United States in an experimental realization of the algorithm that uses a neuromorphic hardware platform comprising phase-change memristive devices. Taken together, our analyses and demonstrations suggest that FSTDP paves the way for the exploitation of the spike-based strengths of SNNs in real-world applications.