From artificial intelligence (AI) and machine learning to CCTV and big data – computer scientists at the University of East Anglia are part of an international effort to make the fishing industry more sustainable. UEA are part of a new £5 million EU-funded project to revolutionise the fishing industry, which employs over 24,000 people in the UK and contributes around £1.4 billion to our economy. It is hoped that pioneering technology will contribute to making the industry more environmentally friendly, sustainable and profitable. The'SMARTFISH-H2020' project, co-ordinated by SINTEF Ocean in Norway, draws on research from universities in Norway, Denmark, Turkey, France and Spain, along with institutes and industry partners across Europe. Other UK partners include Marine Scotland, The Centre for Environment, Fisheries and Aquaculture Science (CEFAS), and Safetynet Technologies Limited.

'Robots are Coming For Our Jobs,' yelled the Huffington Post. 'Robots will destroy our jobs – and we're not ready for it,' The Guardian calmly announced. And according to the Daily Mail, 'Robots taking human jobs is causing a "hellish dystopia"'. These headlines may seem over the top, but, like the scariest nightmares, they're rooted in reality. From factory line jobs to professions like accounting and medicine, artificial intelligence (AI) technologies mean that more and more tasks can now be automated and completed by machines or algorithms. A recent report by think tank Centre for Cities estimates that 1 in 5, or 3.6 million, British jobs will probably be'displaced' by 2030 due to automation and globalisation.

Cloud Robotics is a term that was popularized by James Kuffner after he brought together researchers from different relevant fields (robotics, machine learning, and computer vision) to assist in coming up with the initial Cloud Robotics concept. Cloud robotics, as the name suggests is bringing together cloud computing and robotics. In essence, taking all the benefits of cloud computing and finding ways to apply them to robot software and robotics. The past couple of years have established cloud computing as the technology of now and the future. In 2017, spending on cloud services was $153.5bn, and this is expected to rise by 21.1% in 2018 to $184.4bn.

A two-armed robot in a Chiba factory carefully stacks rice balls in a box, which a worker carries off for shipment to convenience stores. At another food-packaging plant, a robot shakes pepper and powdered cheese over pasta that a person has just arranged in a container. In a country known for bringing large-scale industrial robots to the factory floor, such relatively dainty machines have until recently been dismissed as niche and low-margin. But as the workforce ages in Japan and elsewhere, collaborative robots -- or "cobots" -- are seen as a key way to help keep all types of assembly lines moving without replacing humans. Japan's Fanuc Corp. and Yaskawa Electric Corp., two of the world's largest robot manufacturers, didn't see the shift coming.

The U.S. Army's Autonomous Remote Engagement System is mounted on the Picatinny Lightweight Remote Weapon System and coupled with an M240B machine gun. It's part of a program that reduces the time to identify targets using automatic target detection and user-specified target selection. The U.S. Army's Autonomous Remote Engagement System is mounted on the Picatinny Lightweight Remote Weapon System and coupled with an M240B machine gun. It's part of a program that reduces the time to identify targets using automatic target detection and user-specified target selection. Killer robots have been a staple of TV and movies for decades, from Westworld to The Terminator series. But in the real world, killer robots are officially known as "autonomous weapons."

We introduce techniques for exploring the functionality of a neural network and extracting simple, human-readable approximations to its performance. By performing gradient ascent on the input space of the network, we are able to produce large populations of artificial events which strongly excite a given classifier. By studying the populations of these events, we then directly produce what are essentially contour maps of the network's classification function. Combined with a suite of tools for identifying the input dimensions deemed most important by the network, we can utilize these maps to efficiently interpret the dominant criteria by which the network makes its classification. As a test case, we study networks trained to discriminate supersymmetric stop production in the dilepton channel from Standard Model backgrounds. In the case of a heavy stop decaying to a light neutralino, we find individual neurons with large mutual information with $m_{T2}^{\ell\ell}$, a human-designed variable for optimizing the analysis. The network selects events with significant missing $p_T$ oriented azimuthally away from both leptons, efficiently rejecting $t\overline{t}$ background. In the case of a light stop with three-body decays to $Wb{\widetilde \chi}$ and little phase space, we find neurons that smoothly interpolate between a similar top-rejection strategy and an ISR-tagging strategy allowing for more missing momentum. We also find that a neural network trained on a stealth stop parameter point learns novel angular correlations.

Hyperbolic embeddings offer excellent quality with few dimensions when embedding hierarchical data structures like synonym or type hierarchies. Given a tree, we give a combinatorial construction that embeds the tree in hyperbolic space with arbitrarily low distortion without using optimization. On WordNet, our combinatorial embedding obtains a mean-average-precision of 0.989 with only two dimensions, while Nickel et al.'s recent construction obtains 0.87 using 200 dimensions. We provide upper and lower bounds that allow us to characterize the precision-dimensionality tradeoff inherent in any hyperbolic embedding. To embed general metric spaces, we propose a hyperbolic generalization of multidimensional scaling (h-MDS). We show how to perform exact recovery of hyperbolic points from distances, provide a perturbation analysis, and give a recovery result that allows us to reduce dimensionality. The h-MDS approach offers consistently low distortion even with few dimensions across several datasets. Finally, we extract lessons from the algorithms and theory above to design a PyTorch-based implementation that can handle incomplete information and is scalable.

This paper considers stochastic optimization problems for a large class of objective functions, including convex and continuous submodular. Stochastic proximal gradient methods have been widely used to solve such problems; however, their applicability remains limited when the problem dimension is large and the projection onto a convex set is costly. Instead, stochastic conditional gradient methods are proposed as an alternative solution relying on (i) Approximating gradients via a simple averaging technique requiring a single stochastic gradient evaluation per iteration; (ii) Solving a linear program to compute the descent/ascent direction. The averaging technique reduces the noise of gradient approximations as time progresses, and replacing projection step in proximal methods by a linear program lowers the computational complexity of each iteration. We show that under convexity and smoothness assumptions, our proposed method converges to the optimal objective function value at a sublinear rate of $O(1/t^{1/3})$. Further, for a monotone and continuous DR-submodular function and subject to a general convex body constraint, we prove that our proposed method achieves a $((1-1/e)OPT-\eps)$ guarantee with $O(1/\eps^3)$ stochastic gradient computations. This guarantee matches the known hardness results and closes the gap between deterministic and stochastic continuous submodular maximization. Additionally, we obtain $((1/e)OPT -\eps)$ guarantee after using $O(1/\eps^3)$ stochastic gradients for the case that the objective function is continuous DR-submodular but non-monotone and the constraint set is down-closed. By using stochastic continuous optimization as an interface, we provide the first $(1-1/e)$ tight approximation guarantee for maximizing a monotone but stochastic submodular set function subject to a matroid constraint and $(1/e)$ approximation guarantee for the non-monotone case.

We study rough high-dimensional landscapes in which an increasingly stronger preference for a given configuration emerges. Such energy landscapes arise in glass physics and inference. In particular we focus on random Gaussian functions, and on the spiked-tensor model and generalizations. We thoroughly analyze the statistical properties of the corresponding landscapes and characterize the associated geometrical phase transitions. In order to perform our study, we develop a framework based on the Kac-Rice method that allows to compute the complexity of the landscape, i.e. the logarithm of the typical number of stationary points and their Hessian. This approach generalizes the one used to compute rigorously the annealed complexity of mean-field glass models. We discuss its advantages with respect to previous frameworks, in particular the thermodynamical replica method which is shown to lead to partially incorrect predictions.

Monitoring the biomedical literature for cases of Adverse Drug Reactions (ADRs) is a critically important and time consuming task in pharmacovigilance. The development of computer assisted approaches to aid this process in different forms has been the subject of many recent works. One particular area that has shown promise is the use of Deep Neural Networks, in particular, Convolutional Neural Networks (CNNs), for the detection of ADR relevant sentences. Using token-level convolutions and general purpose word embeddings, this architecture has shown good performance relative to more traditional models as well as Long Short Term Memory (LSTM) models. In this work, we evaluate and compare two different CNN architectures using the ADE corpus. In addition, we show that by de-duplicating the ADR relevant sentences, we can greatly reduce overoptimism in the classification results. Finally, we evaluate the use of word embeddings specifically developed for biomedical text and show that they lead to a better performance in this task.