Inference of Common Multidimensional Equally Distributed Attributes

Given two relations containing multiple measurements – possibly withuncertainties – our objective is to find which sets of attributes from the first have a corresponding set on the second, using exclusively a sample of the data . This approach could be used even when the associated metadata is damaged, missing or incomplete, or when the volume is too big for exact methods .…

Addressing the Vulnerability of NMT in Input Perturbations

Neural Machine Translation (NMT) has achieved significant breakthrough in performance but is known to suffer vulnerability to input perturbations . Asreal input noise is difficult to predict during training, robustness is a big issue for system deployment . In this paper, we improve the robustness of NMT models by reducing the effect of noisy words through a Context-EnhancedReconstruction (CER) approach .…

On Generating and Labeling Network Traffic with Realistic Self Propagating Malware

Synthetically produced data containing fabricated or merged network traffic is of limitedvalue as it is easily distinguishable from real traffic by even simple machinelearning (ML) algorithms . Real-world malware is defanged and introduced to simulated, secured nodes within those networks to generate realistic traffic while maintaining sufficient isolation to protect real data and infrastructure .…

Locally Private k Means in One Round

We provide an approximation algorithm for k-means clustering in the one-round(aka non-interactive) local model of differential privacy (DP) This algorithmachieves an approximation ratio arbitrarily close to the best non-approximation algorithm, improving upon previously known algorithms that only guarantee large (constant) approximation ratios .…

Prospective Artificial Intelligence Approaches for Active Cyber Defence

Cybercriminals are rapidly developing new malicious tools that leverage AI to enable new classes of adaptive and stealthy attacks . New defensive methods need to be developed to counter these threats . Alan Turing Institute, with expert guidance from UK Cyber Security Centre and Defence Science Technology Laboratory, published a research roadmap for AI for ACD last year .…

Inferring Drop in Binary Parsers from Program Executions

We present BIEBER (Byte-IdEntical Binary parsER), the first system to modeland regenerate a working parser from instrumented program executions . Toachieve this, BIEber exploits the regularity (e.g., header fields andarray-like data structures) that is commonly found in file formats . Separate backends (C and Perl in our prototype) can translate the IR intothe same language as the original program (for a safer drop-in replacement), or port to a different language .…

Robust Online Algorithms for Dynamic Choosing Problems

Semi-online algorithms that are allowed to perform a bounded amount of repacking achieve guaranteed good worst-case behaviour in a more realisticsetting . We present a framework for choosing problems that allows usto transfer offline algorithms into $(\alpha-epsilon) algorithms with amortized migration$O(1/\epSilon) We complement these positive results with lower bounds with lowerbounds that show that our results are tight in the sense that no amortization of $o(1)$ is possible .…

Elimination Distance to Topological minor free Graphs is FPT

In the literature on parameterized graph problems, there has been an increased effort in recent years aimed at exploring novel notions of graphedit-distance that are more powerful than the size of a modulator to a specificgraph class . In this paper, we show that deciding whether a given graph has elimination distance at most $k$ to any minor-closed class ofgraphs is fixed-parameter tractable .…

A monolithic and a partitioned Reduced Basis Method for Fluid Structure Interaction problems

The aim of this work is to present a brief report concerning the various aspects of the Reduced Basis Method within Fluid-Structure Interaction problems . The toy problem that we consider is the Turek-Hron benchmark testcase, with a fluid Reynolds number Re = 100, which is known to lead to the formation of Karman vortexes in the fluid, and a periodically oscillatingbehaviour in the structure .…

Understanding Synonymous Referring Expressions via Contrastive Features

Referring expression comprehension aims to localize objects identified by natural language descriptions . Eachobject can be described by synonymous sentences with paraphrases, and suchvarieties in languages have critical impact on learning a comprehension model . We develop an end-to-end trainable framework to learn contrastive features on the image andobject instance levels .…

Neural Tree Expansion for Multi Robot Planning in Non Cooperative Environments

We present a self-improving, neural tree expansion method for multi-robotonline planning in non-cooperative environments . Our algorithm adapts the centralized, perfect information,discrete-action space method from Alpha Zero to a decentralized, partial information, continuous action space setting . Our numerical experiments demonstrate neural expansion generatescompact search trees with better solution quality and 20 times less computing expense compared to MCTS without neural expansion .…

Towards Autonomous Robotic Precision Harvesting

A walking harvester performs the challenging task of autonomous navigation and tree grabbing in a confined, GPS denied forestenvironment . Strategies for mapping, localization, planning, and control are proposed and integrated into a fully autonomous system . The mission starts with a human or a mobile robot mapping the area of interest using a custom-madesensor module .…

The Emperor s New Autofill Framework A Security Analysis of Autofill on iOS and Android

Password managers help users more effectively manage their passwords, encouraging them to adopt stronger passwords across their many accounts . Mobile operating systems provide autofill frameworks that are designed to integrate with password managers . They also enforce insecure behavior and fail toprovide the password managers implemented using the frameworks with sufficient information to override this incorrect behavior .…

Solution landscape of Onsager functional identifies non axisymmetric critical points

We investigate the solution landscapes of the Onsager free-energy functional with different potential kernels . For the coupled dipolar/Maier–Saupe potential, the solutionlandscape shows a novel non-axisymmetric critical point, named as tennis, whichexists for a wide range of parameters . The solution landscape provides an efficient approach to show the global structures as well as the bifurcations of critical points, which can not only verify the previous analytic results but also proposes conjectures based on the numerical findings .…

Market Value of Differentially Private Smart Meter Data

This paper proposes a framework to investigate the value of sharingprivacy-protected smart meter data between domestic consumers and load servingentities . The framework consists of a discounted differential privacy model toensure individuals cannot be identified from aggregated data . It also includes a ANN-basedshort-term load forecasting to quantify the impact of data availability and privacy protection on the forecasting error and an optimal procurement day-ahead and balancing markets to assess the market value of the privacy-utility trade-off .…

Does enhanced shape bias improve neural network robustness to common corruptions

Convolutional neural networks learn to extract representations of complex features such as object shapes and textures to solve image recognition tasks . Recent work indicates CNNs trained on ImageNet are biased towardsfeatures that encode textures . These alone are sufficient to generalizeto unseen test data from the same distribution as the training data but often fail to generalize to out-of-distribution data .…

Flow based Video Segmentation for Human Head and Shoulders

Flow-based encoder-decodernetwork (FUNet) combines traditional Horn-Schunck optical-flowestimation technique and convolutional neural networks to perform robustreal-time video segmentation . Code and pre-trained models are available on GitHub repository:\url{https://://://www.g.com/kuangzijian/Flow-Based-based-Video-Matting.net . We also introduce a video and image segmentationdataset: ConferenceVideoSegmentationDataset .…

Deep Learning based Efficient Symbol Level Precoding Design for MU MISO Systems

The recently emerged symbol-level precoding (SLP) technique has been regarded as a promising solution in multi-user wireless communication systems . However, the tremendous computational complexity of conventional SLP designs severely hinders the practical implementation of the technique . The proposed EPNN based SLP design candramatically reduce the computing time at the price of slight performance loss .…

On the social and cognitive dimensions of wicked environmental problems characterized by conceptual and solution uncertainty

Wicked problems reflect our incomplete understanding of interdependent global systems and the hyper-risk they pose . Such problems escape solutions because they are often ill-defined and thus mis-identified andunder-appreciated by problem-solvers and the communities they constitute . We develop a quantitative framework for understanding the class of wicked problems that emerge at the intersections of natural, social, and technological complex systems .…

HMS Hierarchical Modality Selectionfor Efficient Video Recognition

Conventional video recognition pipelinestypically fuse multimodal features for improved performance . But this is not only computationally expensive but also neglects the fact that differentvideos rely on different modalities for predictions . HMS operates on a low-costmodality, i.e., audio clues, by default, and dynamically decides on-the-fly whether to use computationally-expensive modalities on a per-input basis .…

Detector Free Weakly Supervised Grounding by Separation

Weakly Supervisedphrase-Grounding (WSG) deals with the task of using this data to learn to localize (or to ground) arbitrary text phrases in images without any additionalannotations . Most recent SotA methods for WSG assume the existence of a pre-trained object detector, relying on it to produce the ROIs for localization .…

Review of end to end speech synthesis technology based on deep learning

Speechsynthesis technology helps users get the output of intelligent machine more easily and intuitively . Current research focus is the deep learning-basedend-to-end speech synthesis technology . It mainly consists of three modules: textfront-end, acoustic model, and vocoder . This paper reviews the research status of these three parts, and classifies and compares various methods according to their emphasis .…

Towards Solving Multimodal Comprehension

This paper targets the problem of procedural multimodal machine comprehension(M3C) This task requires an AI to comprehend given steps of multimodalinstructions and then answer questions . Yagcioglu etal. [35] introduced RecipeQA dataset to evaluate M3C . We hypothesized that naturally occurring bias present in the dataset affects even the bestperforming model .…

CTNet Context based Tandem Network for Semantic Segmentation

Spatial Contextual Module (SCM) leveraged to learn spatial contextual dependency between pixels by exploring thecorrelation between pixels and categories . The Channel ContextualModule (CCM) is introduced to learn the semantic features including thesemantic feature maps and class-specific features . The learned semantic features are used as the prior knowledge to guide the learning of SCM, which can makeSCM obtain more accurate long-range spatial dependency .…

Finding Small Multi Demand Set Covers with Ubiquitous Elements and Large Sets is Fixed Parameter Tractable

We study a variant of Set Cover where each element of the universe has somedemand that determines how many times the element needs to be covered . We prove that all three problems are fixed-parameter tractable with respect to the combinedparameter budget, the maximum number of elements missing in one of the sets, and how many sets can be included multiple times .…

Multilevel Polar Coded Space Shift Keying

Multilevel coding (MLC) is a coded modulation technique which can achieve excellent performance over a range of communication channels . Polar codes have been shown to be quite compatible with communication systems using MLC, as therate allocation of the component polar codes follows the natural polarizationinherent in polar codes .…

Robust MMSE Precoding and Power Allocation for Cell Free Networks

An iterative robust minimum mean-squareerror (RMMSE) precoder based on generalized loading is developed to mitigate interference in the presence of imperfect channel state information (CSI) Anachievable rate analysis is carried out and optimal and uniform powerallocation schemes are developed based on the signal-to-interference-plus-noiseratio .…