Businesses need to understand AI before putting it to work


Artificial intelligence is a big deal for business – it's the biggest marketing buzzword this side of cryptocurrency and it's set to make a very real difference to the future of work. But what does that really look like? How will AI support employees? And what responsibilities do you have when you start applying AI to customer data? BT's Adastral Park research facility is at the forefront of the development and application of AI technologies in the UK, from using real-time network analysis to detect and protect against cyberattacks as they happen, to enabling customer service agents to anticipate the future needs of their clients based on customer behaviour trends.

AWS Global Accelerator to boost performance across regions


Amazon Web Services (AWS) has lifted the lid off a handful of products and services ahead of its annual re:Invent conference in Las Vegas, announcing on Monday night a network service that allows for the automatic routing of traffic to multiple regions. The AWS Global Accelerator has been touted by VP of global infrastructure at AWS Peter DeSantis as improving availability and performance for AWS customers' end users. Essentially, user traffic enters AWS Global Accelerator through the closest edge location. The accelerator then routes the user traffic to the closest healthy application endpoint within the global AWS network. Lastly, at the endpoint, the application response returns over the AWS global network and reaches the user through the optimal endpoint.

Black Friday 2018: NordVPN is offering 75% off a 3-year VPN plan


Black Friday isn't just for retail and electronics, kids-- NordVPN is kicking off their Cyber Month with a killer Black Friday deal that'll get you a 3-year VPN plan for just $2.99/month. Ever since the FCC decided to kill off Net Neutrality back in December 2017, protecting online privacy has never been more crucial. This is where VPNs come in. As NordVPN puts it, "VPN is essentially a hack-proof, encrypted tunnel for online traffic to flow." In other words, nobody besides you has access to your internet data--meaning you can stream, download, and search for things without worrying about hackers or unwanted third parties accessing your private information, even when traveling or using public Wi-Fi.

SK Telecom and Samsung to collaborate on 5G for enterprise


SK Telecom and Samsung will collaborate in 5G research and development as South Korea gears up for the imminent rollout of the next-generation network. Specifically, the two will collaborate in mobility enhancement for 3.5GHz and 28GHz millimetre-wave (mmWave) spectrum bands, widen coverage, and enhance 5G use cases. South Korean telecommunication carriers are expected to deploy 5G for consumers in March next year, but enterprise clients will start using the network in trials next month. The two companies will build tech to utilise both, in order to increase data transfer rates and distances. The two will also build tech to widen the coverage of 28GHz mmWave spectrum.

Is this the next big spectrum block for IoT? - Stacey on IoT Internet of Things news and analysis


This week, the Federal Communications Commission begins an auction for two chunks of radio waves that have long been considered useless. Thanks to new technologies and demand for wireless broadband, the market is ready to buy 24GHz and 28GHz spectrum. Meanwhile, in the unlicensed band, which anyone can use and isn't sold off by the FCC, another of the so-called millimeter wave spectrum bands is gaining interest. Roughly a decade ago, engineers had hoped to use the 60GHz band for sending fat files over short distances using a technology called ultra-wideband. At the time, the market didn't see a compelling need for the technology, so it was shelved.

Samsung to invest $22 billion in 5G and AI


Samsung Electronics will invest $22 billion in 5G networking and AI going forward to secure a "minimum" of 20 percent market share in network equipment by 2020, the company's network boss has said. Youngky Kim, president and head of Samsung's network business, speaking at WSJ D.Live in California, told a panel that the next-generation network will unlock the potential of artificial intelligence (AI), describing 5G as "oxygen" for AI. "AI needs a lot of data to respond to you," Kim said. "This amount of data can be provided by 5G, not 4G." Samsung produces half a billion electronic devices yearly, and this will provide it with international experience of what humans want, he said.

Multivariate Spatiotemporal Hawkes Processes and Network Reconstruction Machine Learning

There is often latent network structure in spatial and temporal data and the tools of network analysis can yield fascinating insights into such data. In this paper, we develop a nonparametric method for network reconstruction from spatiotemporal data sets using multivariate Hawkes processes. In contrast to prior work on network reconstruction with point-process models, which has often focused on exclusively temporal information, our approach uses both temporal and spatial information and does not assume a specific parametric form of network dynamics. This leads to an effective way of recovering an underlying network. We illustrate our approach using both synthetic networks and networks constructed from real-world data sets (a location-based social media network, a narrative of crime events, and violent gang crimes). Our results demonstrate that, in comparison to using only temporal data, our spatiotemporal approach yields improved network reconstruction, providing a basis for meaningful subsequent analysis --- such as community structure and motif analysis --- of the reconstructed networks.

Stream Reasoning in Temporal Datalog Artificial Intelligence

In recent years, there has been an increasing interest in extending traditional stream processing engines with logical, rule-based, reasoning capabilities. This poses significant theoretical and practical challenges since rules can derive new information and propagate it both towards past and future time points; as a result, streamed query answers can depend on data that has not yet been received, as well as on data that arrived far in the past. Stream reasoning algorithms, however, must be able to stream out query answers as soon as possible, and can only keep a limited number of previous input facts in memory. In this paper, we propose novel reasoning problems to deal with these challenges, and study their computational properties on Datalog extended with a temporal sort and the successor function (a core rule-based language for stream reasoning applications).

Space Communication with Neural Network Resource Allocation


Space may be the final frontier, but it continues to pose myriad technical challenges as commercial and government-driven space investment continues. One of those challenges is developing more effective space-based communication systems for the increasing number of satellites and spacecrafts that need to interact with one another in the void. A team of researchers has developed an algorithm to enable cognitive radio functions on satellite communications systems to adapt themselves autonomously. Current space communication systems deploy radio-resource selection algorithms, but they are rudimentary and work with a pre-programmed look-up table. Furthermore, they have little flexibility regarding the various parameters for the performance goals the system needs to achieve.

Drop-Activation: Implicit Parameter Reduction and Harmonic Regularization Machine Learning

Overfitting frequently occurs in deep learning. In this paper, we propose a novel regularization method called Drop-Activation to reduce overfitting and improve generalization. The key idea is to \emph{drop} nonlinear activation functions by setting them to be identity functions randomly during training time. During testing, we use a deterministic network with a new activation function to encode the average effect of dropping activations randomly. Experimental results on CIFAR-10, CIFAR-100, SVHN, and EMNIST show that Drop-Activation generally improves the performance of popular neural network architectures. Furthermore, unlike dropout, as a regularizer Drop-Activation can be used in harmony with standard training and regularization techniques such as Batch Normalization and AutoAug. Our theoretical analyses support the regularization effect of Drop-Activation as implicit parameter reduction and its capability to be used together with Batch Normalization.