Search

The existing plethora of document classification techniques exploits different data sources either from the content or metadata of research articles. Various journal publishers like Springer, Elsevier, IEEE, etc., do not provide open access to the content of research articles, whereas metadata is freely available there. Metadata like title, keyword, and abstract can serve as a better alternative to the content in various scenarios. In the current literature, researchers have assessed the role of some of the metadata individually. We believe that the collective contribution of metadata parameters can play a significant role in classifying research papers. This paper presents a comprehensive evaluation of the role of metadata, individually as well as in combinations to achieve the objective of research paper classification. Moreover, we have classified the research articles into ACM hierarchy root categories (e.g. general literature, hardware, software, etc.). In this comprehensive evaluation, we have assessed all the possible combinations of metadata features against different classifiers such as Random Forest, K Nearest Neighbor, and Decision Tree. The results of this research reveal that the title & keywords combination outperforms other combinations with an F-measure score of 0.88.

Paper-based digital microfluidic biochips (or P-DMFBs) are becoming highly impelling due to its low-cost and in-place fabrication of electrodes and control wiring on a single piece of paper having an inkjet printer and conductive ink. Despite enormous advantages, several complex design rules also subsist, such as avoidance of induced control interference, minimum separation among the control lines, and congestion-free wiring on a single layer, which is to be correlated leading toward overall feasibility of the design. Several cost raising issues, such as schedule length, control pin count, and wire length, must be considered for attaining a successful fluid-control codesign. Moreover, design gaps exist among the subtasks of the fluid level, control level, and fluid-control design as a whole, which undeniably impose expensive design cycles increasing overall cost. This article builds a machine learning-based model for the pin-constrained P-DMFBs to predict violation in control design beforehand and accordingly guides the fluid-control codesign to tackle important cost-driving issues while attaining congestion- and conflict-free wiring. This model effectively eliminates the design cycles producing a low-cost platform. The predictive model has been evaluated over a balanced data set. Several benchmarks for assessing the performance are studied.

Due to the increasing size of s (s), their design and optimization phases (i.e., ) grow increasingly complex. At design time, a large design space needs to be explored to find an implementation that fulfills all specifications and then optimizes metrics like energy, area, delay, reliability, etc. At run time, a large configuration space needs to be searched to find the best set of parameters (e.g., voltage/frequency) to further optimize the system. Both spaces are infeasible for exhaustive search typically leading to heuristic optimization algorithms that find some trade-off between design quality and computational overhead. ML can build powerful models that have successfully been employed in related domains. In this survey, we categorize how () may be used and is used for design-time and run-time optimization and exploration strategies of s. A meta-study of published techniques unveils areas in that are well-explored and underexplored with, as well as trends in the employed algorithms. We present a comprehensive categorization and summary of the state of the art on for. Finally, we summarize remaining challenges and promising open research directions.

In this article, we put forward the substantial challenges in cyber resilience in the domain of autonomous systems and outline foundational solutions to address these challenges. These solutions fall into two broad themes: resilience-by-design and resilience-by-reaction . We use several application drivers from autonomous systems to motivate the challenges in cyber resilience and to demonstrate the benefit of the solutions. We focus on some autonomous systems in the near horizon (autonomous ground and aerial vehicles) and also a little more distant (autonomous rescue and relief). For resilience-by-design , we focus on design methods in software that are needed for our cyber systems to be resilient. In contrast, for resilience-by-reaction , we discuss how to make systems resilient by responding, reconfiguring, or recovering at runtime when failures happen. We also discuss the notion of adaptive execution to improve resilience, execution transparently and adaptively among available execution platforms (mobile/embedded, edge, and cloud). For each of the two themes, we survey the current state, and the desired state and ways to get there. We conclude the paper by looking at the research challenges we will have to solve in the short and the mid-term to make the vision of resilient autonomous systems a reality. This article came out of discussions that started at the NSF-sponsored Grand Challenges in Resilience Workshop held at Purdue in 2019 with the co-authors contributing to going into the depth of the issues and then this article.

Introduction: The electrocardiogram (ECG) plays an important role in the diagnosis of heart diseases. However, most patterns of diseases are based on old datasets and stepwise algorithms that provide limited accuracy. Improving diagnostic accuracy of the ECG can be done by applying machine learning algorithms. This requires taking existing scanned or printed ECGs of old cohorts and transforming the ECG signal to the raw digital (time (milliseconds), voltage (millivolts)) form. Objectives: We present a MATLAB-based tool and algorithm that converts a printed or scanned format of the ECG into a digitized ECG signal. Methods: 30 ECG scanned curves are utilized in our study. An image processing method is first implemented for detecting the ECG regions of interest and extracting the ECG signals. It is followed by serial steps that digitize and validate the results. Results: The validation demonstrates very high correlation values of several standard ECG parameters: PR interval 0.984 +/−0.021 (p-value < 0.001), QRS interval 1+/− SD (p-value < 0.001), QT interval 0.981 +/− 0.023 p-value < 0.001, and RR interval 1 +/− 0.001 p-value < 0.001. Conclusion: Digitized ECG signals from existing paper or scanned ECGs can be obtained with more than 95% of precision. This makes it possible to utilize historic ECG signals in machine learning algorithms to identify patterns of heart diseases and aid in the diagnostic and prognostic evaluation of patients with cardiovascular disease., We present a MATLAB-based tool and algorithm that converts a printed or scanned ECG into a digitized ECG signal. Validation on 30 ECG scanned curves more than 95% precision. This makes it possible to utilize historic ECG signals in machine learning algorithms to identify patterns of heart diseases and aid in the diagnosis and evaluation of patients with cardiovascular disease.

Designing clean, reusable, and repeatable experiments for a research paper does not have to be difficult. We report on our efforts to create an integrated toolchain for running, processing, and including the results of computer experiments in scientific publications.

Transmission of video traffic over the Internet has grown exponentially in the past few years with no sign of waning. This increasing demand for video services has changed user expectation of quality. Various mechanisms have been proposed to optimise the Quality of Experience (QoE) of end users video. Studying these approaches are necessary for new methods to be proposed or combination of existing ones to be tailored. We discuss challenges facing the optimisation of QoE for video traffic in this paper. It surveys and classifies these mechanisms based on their functions. The limitation of each of them is identified and future directions are highlighted., 18 pages, 7 figures

THE YEAR IS 2028. IT'S 8 P.M. ON A Wednesday night and you're famished. You're staring wistfully at the only remaining item in your refrigerator: a package of sausages with an unappetizing grayish hue. Ugh. Did they always look like that? Are they still safe to eat? In 2018, you'd have to rely on your sense of smell and take a gamble. But in 2028, you might simply wave your smartphone over the package. The smartphone interrogates the package's embedded sensor, which measures the concentration of gases associated with meat decomposition. The smartphone displays the message "Safe to eat within the next 20 hours," and then offers a list of recipes for cooking with sausages. Too hungry to bother with the recipes, you tear open the package, toss the sausages into a frying pan, and discard the package-along with its sensor technology.

Digital pen and paper technologies are proving increasingly popular in vertical markets such as health care and manufacturing, which make heavy use of standardized forms that can easily be produced on specialized paper. But the still-untapped frontier is the horizontal zone of everyday consumers. Within this vast and tempting territory resides the commercial digital pen's most-often targeted subgroup: notetakers - whether they're journalists, lawyers, or the many students who sit daily with their heads bobbing between their paper or computer notebooks and the lecturer at the podium. In the past year, several new digital pens have targeted this group in the hope of uniting the broader market's paper and digital worlds. And one digital pen - Livescribe's Pulse smartpen - is working out its grand vision of a paper-based mobile computing platform that it hopes will radically alter how we view pens and their possibilities.

The eminent growth in connected devices technologies, like the Internet of Things (IoT), 5 th generation (5G) and beyond communication systems, lead to innovations for human beings. The uninterrupted data communication services play diverse roles in the mobile user's routine life. The various changing user group patterns are unpredictable due to the day-by-day up-gradation of user devices. Even though the existing telecommunication services have upgraded up to 5G with high performance distributed computing (HPDC) in the distributed data networks (DDN) for ultra-reliable and low latency communication (URLLC) services, yet the backend distributed database systems (DDS) are facing serious security and privacy issues due to lack of federated intelligence in the data networks. Recently, we have seen an increased focus and effort by users and policymakers toward enhancing security and privacy related to the collection and usage of the data in DDNs. When it comes to the intelligence using machine/deep learning (ML/DL) of the HPDC systems for security and privacy, enough dataset is required, which often includes personal user information to train ML/DL models. As data privacy and security represents a growing critical concern, given the above-mentioned new areas of legislation and policies, novel ML methodologies like federated learning (FL) have been developed in part to address these concerns.

The objective of this paper was to provide an automatic engine to classify and locate information using natural language. The proposal integrates a set of two algorithms to extract information from different repositories using their own open APIs and creates a knowledge database using a natural language approach using a Bayesian algorithm to classify and a second algorithm to clean the paper. Putting said techniques together derived in a strong alternative which reach common gaps in classification and location of information including avoid the use of the whole paper to get information and not only the information introduced at the moment of upload the paper in the digital library. The proposal was oriented to classify and locate research papers in order to better describe this contribution, however, findings could be applicable to a vast range of scenarios. An adaptation of the popular methodology Crisp-DM was used to evaluate the performance of the algorithm obtaining good results in classifying, searching, and feeding the knowledge base.

This special issue is the fourth highlighting selected papers from the IEEE Symposium on Security and Privacy. Several articles presented address vulnerabilities, and possible mitigation strategies, related to smart devices, home systems, entertainment systems, and smartphones, while others study human behavior in managing devices, security- and privacy-relevant code development, and critical infrastructure management. Given the intense discussions in the Internet of Things (IoT) community and the increasing complexity of human-system interactions, these are timely topics.

In this paper, we investigate a very challenging task of automatically generating presentation slides for academic papers. The generated presentation slides can be used as drafts to help the presenters prepare their formal slides in a quicker way. A novel system called PPSGen is proposed to address this task. It first employs the regression method to learn the importance scores of the sentences in an academic paper, and then exploits the integer linear programming (ILP) method to generate well-structured slides by selecting and aligning key phrases and sentences. Evaluation results on a test set of 200 pairs of papers and slides collected on the web demonstrate that our proposed PPSGen system can generate slides with better quality. A user study is also illustrated to show that PPSGen has a few evident advantages over baseline methods.

Writer verification is a method to specify an authentic writer from handwriting. Automated writer verification methods are required for various applications (e.g., credit cards, checks, and passports). However, there is room for improvement in the performance of such methods compared with the performance of human beings, for example, forensic document examiners. Because automated writer verification systems do not always return correct results under any circumstances, which can lead to grave consequences, further research is required to improve the performance of such methods. Furthermore, problems caused by limited samples must be solved for real applications. To improve verification accuracy with limited samples, we propose a text and user generic model for writer verification that uses a combination of pen pressure information from ink intensity and writing indentations obtained by a multiband image scanner. We introduce a writer-specific dissimilarity representation to consider individual handwriting characteristics that affect model performance. Experimental results obtained using handwriting samples collected from 54 volunteers are reported. The results show a decrease in error rate compared with conventional methods from 10.0% to 4.0%.

Visualizing the intellectual structure of scientific domains using co-cited units such as references or authors has become a routine for domain analysis. In previous studies, paper-reference matrices are usually transformed into reference-reference matrices to obtain co-citation relationships, which are then visualized in different representations, typically as node-link networks, to represent the intellectual structures of scientific domains. Such network visualizations sometimes contain tightly knit components, which make visual analysis of the intellectual structure a challenging task. In this study, we propose a new approach to reveal co-citation relationships. Instead of using a reference-reference matrix, we directly use the original paper-reference matrix as the information source, and transform the paper-reference matrix into an FP-tree and visualize it in a Java-based prototype system. We demonstrate the usefulness of our approach through visual analyses of the intellectual structure of two domains: Information Visualization and Sloan Digital Sky Survey (SDSS). The results show that our visualization not only retains the major information of co-citation relationships, but also reveals more detailed sub-structures of tightly knit clusters than a conventional node-link network visualization.

This paper presents CoScribe, a concept and prototype system for the combined work with printed and digital documents, which supports a large variety of knowledge work settings. It integrates novel pen-and-paper-based interaction techniques that enable users to collaboratively annotate, link and tag both printed and digital documents. CoScribe provides for a very seamless integration of paper with the digital world, as the same digital pen and the same interactions can be used on paper and displays. As our second contribution, we present empirical results of three field studies on learning at universities. These motivated the design of CoScribe and were abstracted to a generic framework for the design of intuitive pen-and-paper user interfaces. The resulting interaction design comprising collaboration support and multiuser visualizations has been implemented and evaluated in user studies. The results indicate that CoScribe imposes only minimal overhead on traditional annotation processes and provides for a more efficient structuring and retrieval of documents.

This issue of the Communications Standards Supplement includes updated versions of selected papers from the seventh International Telecommunications Union (ITU) Kaleidoscope Academic Conference. The conference took place in Barcelona, Spain, 9-11 December 2015, as part of the celebration of the ITU's 150th anniversary. The topic of the conference was "Trust in the Information Society", and the event was technically co-sponsored by the Institute of Electrical and Electronics Engineers (IEEE), the IEEE Communication Society, and the Institute of Electronics, Information and Communication Engineers of Japan (IEICE). The conference host was the Universitat Autonoma de Barcelona, Spain.

When faced with the tasks of reading and writing a complex technical paper, many nonnative scientists and engineers who have a solid background in English grammar and vocabulary lack an adequate knowledge of commonly used structural patterns at the discourse level. In this paper, we propose a novel computer software tool that can assist these people in the understanding and construction of technical papers, by automatically identifying the structure of writing in different fields and disciplines. The system is tested using research article abstracts and is shown to be a fast, accurate, and useful aid in the reading and writing process.

Wireless communication is continuing to make inroads into many facets of society and is gradually becoming more and more ubiquitous. While in the past wireless communication (as well as mobility) was largely limited to the first and last transmission hops, today's wireless networks are starting to offer purely wireless, often mobile, and even opportunistically connected operation. The purpose of this article is to examine security and privacy issues in some new and emerging types of wireless networks, and attempt to identify directions for future research.

Since the origins of science, researchers' main output has been books and papers that describe new theoretical and practical ideas and results. These sources are useful and important, but learning from them is often not easy. How can we facilitate this learning process?

Personal computers (PCs) are increasingly being used in industrial applications. Graphical user interface (GUI) software and Windows-based operating systems are proving to be cost-effective, rugged and reliable. This paper describes how these advances have, over the past four years, allowed the successful installation of several human-machine interface (HMI) systems for paper machine winders.

Previously, companies could publish the interfaces or other aspects of their products and expect that they would become a standard. In fact, there are more examples of this failing, than examples where a company's market position made its product the de facto standard. Today, consumers and customers expect multi-vendor interoperability and will not tolerate proprietary interfaces. There is a need for market driven standards.

This Series presents selected papers from the ITU Kaleidoscope Academic Conference that convened in Kyoto, Japan from the 22nd to 24th of April 2013. The topic of the conference was "Building Sustainable Communities," in recognition of the challenges that Japan is facing after the Great East Earthquake. The hosts were the Ministry of Internal Affairs and Communication (MIC) of Japan and Kyoto University. Kyoto University is the second oldest Japanese national university after the University of Tokyo. It was founded in 1897 following the Meiji Restoration, which adopted various western systems to build a modern state.

The principles of the model-based document analysis system called Pi ODA (paper interface to office document architecture), which was developed as a prototype for the analysis of single-sided business letters in German, are presented. Initially, Pi ODA extracts a part-of hierarchy of nested layout objects such as text-blocks, lines, and words based on their presentation on the page. Subsequently, in a step called logical labeling, the layout objects and their compositions are geometrically analyzed to identify corresponding logical objects that can be related to a human perceptible meaning, such as sender, recipient, and date in a letter. A context-sensitive text recognition for logical objects is then applied using logical vocabularies and syntactic knowledge. As a result, Pi ODA produces a document representation that conforms to the ODA international standard. >

The three articles in this special section were selected as the "Best Papers" presented at IEEE WiVeC 2013.

The Guest Editor summarizes the three papers in this special section. These papers were presented at the Annual International Conference on Mobile Computing and Networking (MobiCom) held in August 2012 in Istanbul, Turkey.

In Part II of this series, I went over how to make life easy for the IEEE MTT-S International Microwave Symposium (IMS) reviewer so that we can increase the chances of getting a paper accepted. It is important to devote conscious effort into making it easy for the reviewer to give high scores in the originality, clarity, quantitative, and interest categories. In this part, I concentrate on what I feel is the most important category, quantitative. Note that all four categories are formally assigned equal weights when assigning scores, and other reviewers might have different opinions, so this represents only my personal views on the matter.

Objective: We present a Matlab-based tool to convert electrocardiography (ECG) information from paper charts into digital ECG signals. The tool can be used for long-term retrospective studies of cardiac patients to study the evolving features with prognostic value. Methods and procedures: To perform the conversion, we: 1) detect the graphical grid on ECG charts using grayscale thresholding; 2) digitize the ECG signal based on its contour using a column-wise pixel scan; and 3) use template-based optical character recognition to extract patient demographic information from the paper ECG in order to interface the data with the patients' medical record. To validate the digitization technique: 1) correlation between the digital signals and signals digitized from paper ECG are performed and 2) clinically significant ECG parameters are measured and compared from both the paper-based ECG signals and the digitized ECG. Results: The validation demonstrates a correlation value of 0.85–0.9 between the digital ECG signal and the signal digitized from the paper ECG. There is a high correlation in the clinical parameters between the ECG information from the paper charts and digitized signal, with intra-observer and inter-observer correlations of 0.8–0.9 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$({\rm p}

This special issue includes three highest ranked papers selected by the TPC chairs and members from the technical track of the 2011 DySPAN conference on dynamic spectrum access.

A Canadian feasibility study shows that the key to reliable and cost-effective automatic reading of drawings is detailed planning of an operator-assisted system.

A comparison between the models of scientific communication and mass communication is made in terms of the basic components in the communication processes. Differences between the two models are emphasized so that the sources of ineffective communication can be identified. Several suggestions are made to bridge the communication gaps between science, technology, and the public.

The difficulty of very high resolution image change detection lies in the low interclass separability between the changed class and the unchanged class. According to experiments, we found that this separability can be improved by mining the relationship contained in the training samples. Based on this observation, a supervised change detection approach is proposed in this paper based on relationship learning. The proposed approach begins with enriching the training samples based on their neighborhood relationship and label coherence; this relationship is then learned simultaneously with the classifier, and, finally, the latter classification performance benefits from the learned relationship. Experiments demonstrate the effectiveness of the proposed approach.

Here it comes, September, time for new plans. In August you received an e-mail announcing the new online version of the Magazine offered, as well as the print version, to AESS members. Actually, until the end of the year the online version of the Magazine is offered free to the public. In 2017, this offering will be a benefit of AESS membership. The online version has not been created to replace the print version, because we know that many of you like to have the paper version to read and to bring everywhere. But, sometimes, it's easier and faster to browse the online version on a laptop or tablet. That's why we have decided to offer the online version of the Magazine as well. The idea for the future is not to have an exact replica of the print version. The content will be the same, but on the online version we can add new features associated to the papers, like videos, active links and so on. If you have suggestions or simply comments, please contact me or our administrative editor William Walsh (w.walsh@ieee.org). All ideas, proposals, and feedback are welcome. Now I leave my space to the guest editorial of the special issue on Sense & Avoid, Part II. Enjoy reading and ciao.

Digital video now plays an important role in supporting more profitable online patient training and counseling, and integration of patient training videos from multiple competitive organizations in the health care network will result in better offerings for patients. However, privacy concerns often prevent multiple competitive organizations from sharing and integrating their patient training videos. In addition, patients with infectious or chronic diseases may not want the online patient training organizations to identify who they are or even which video clips they are interested in. Thus, there is an urgent need to develop more effective techniques to protect both video content privacy and access privacy . In this paper, we have developed a new approach to construct a distributed Hippocratic video database system for supporting more profitable online patient training and counseling. First, a new database modeling approach is developed to support concept-oriented video database organization and assign a degree of privacy of the video content for each database level automatically. Second, a new algorithm is developed to protect the video content privacy at the level of individual video clip by filtering out the privacy-sensitive human objects automatically. In order to integrate the patient training videos from multiple competitive organizations for constructing a centralized video database indexing structure, a privacy-preserving video sharing scheme is developed to support privacy-preserving distributed classifier training and prevent the statistical inferences from the videos that are shared for cross-validation of video classifiers. Our experiments on large-scale video databases have also provided very convincing results.

Computing with words (CWW) means different things to different people. This article is the start of a position paper, written by some of the members of the CIS fuzzy systems technical committee task force on CWW, that answers the question "What does CWW mean to me"?

This issue's Works in Progress department includes five submissions covering a variety of research topics. The first two projects are sensor network oriented: a Java MIDP-based sensor network platform that's being offered as an open source development environment and a platform for monitoring and controlling distributed power generation systems. The third is a wearable computing application for optimizing the circuit-training process for a group of users. The final two projects introduce pervasive computing into medical environments with a combined electronic and paper medical charting system and an interactive, multimodal, haptic-enabled medical training simulator.

Researchers have begun developing a color version of electronic paper, just when the monochromatic version has begun to take off in consumer applications. E Ink is one company working on the concept and has already released a demonstration version.

Looks at efforts now under way to resolve the issue of copyright infringement liability for on-line Internet access service providers. First, the Administration's Information Infrastructure Task Force (IITF) released a White Paper report recommending that present liability for copyright infringement be kept in place for Internet purposes, without change, because that would give Net access service providers an incentive to prevent and avoid infringement. Congress then introduced two bills, implementing the IITF proposals. These bills simply provided, as the IITF had recommended, that unauthorized transmission of copyrighted material between computers or over the Net was "distribution" and thus copyright infringement. The author looks at the ensuing uproar, and its resolution.

Inhomogeneous linear ordinary differential equations (ODEs) and systems of ODEs can be solved in a variety of ways. However, hardware circuits that can perform the efficient analog computation to solve them are rarely in the literature. To address such problems, this paper proposes a general method of using a memristor-capacitor (M-C) circuit to solve inhomogeneous linear ODEs and systems of ODEs of any order in initial value problems. The M-C circuit can match the coefficients of the equations sought by adjusting the memristor resistance value according to the coefficient formula proposed in the paper, which has higher programmability. Then, some ODEs and systems of ODEs are given in the paper as examples to evaluate the proposed method. According to the comparison results based on MATLAB software simulation and the simulation based on OrCAD software, the designed M-C circuit has an effective improvement in speed and the accuracy exceeds 99.95% in software simulation. Based on practical verification, this paper gives the actual M-C circuit experiment based on PCB. Moreover, the proposed method can be used to quickly solve the object motion state in the spring mass damping system in actual engineering, and the accuracy can reach 99.98%.

This paper aims to predict a set of apps a user will open on her mobile device in the next time slot. Such an information is essential for many smartphone operations, e.g., app pre-loading and content pre-caching, to improve user experience. However, it is hard to build an explicit model that accurately captures the complex environment context and predicts a set of apps at one time. This paper presents a deep reinforcement learning framework, named as DeepAPP, which learns a model-free predictive neural network from historical app usage data. Meanwhile, an online updating strategy is designed to adapt the predictive network to the time-varying app usage behavior. To transform DeepAPP into a practical deep reinforcement learning system, several challenges are addressed by developing a context representation method for complex contextual environment, a general agent for overcoming data sparsity and a lightweight personalized agent for minimizing the prediction time. Extensive experiments on a large-scale anonymized app usage dataset reveal that DeepAPP provides high accuracy (precision 70.6% and recall of 62.4%) and reduces the prediction time of the state-of-the-art by 6.58 times. A field experiment of 29 participants demonstrates DeepAPP can effectively reduce launch time of apps.

Available data in machine learning applications is becoming increasingly complex, due to higher dimensionality and difficult classes. There exists a wide variety of approaches to measuring complexity of labeled data, according to class overlap, separability or boundary shapes, as well as group morphology. Many techniques can transform the data in order to find better features, but few focus on specifically reducing data complexity. Most data transformation methods mainly treat the dimensionality aspect, leaving aside the available information within class labels which can be useful when classes are somehow complex. This paper proposes an autoencoder-based approach to complexity reduction, using class labels in order to inform the loss function about the adequacy of the generated variables. This leads to three different new feature learners, Scorer, Skaler and Slicer. They are based on Fisher's discriminant ratio, the Kullback-Leibler divergence and least-squares support vector machines, respectively. They can be applied as a preprocessing stage for a binary classification problem. A thorough experimentation across a collection of 27 datasets and a range of complexity and classification metrics shows that class-informed autoencoders perform better than 4 other popular unsupervised feature extraction techniques, especially when the final objective is using the data for a classification task., This paper has been accepted for publication by IEEE Transactions on Pattern Analysis and Machine Intelligence

Deep learning models have been shown to be vulnerable to adversarial attacks. Adversarial attacks are imperceptible perturbations added to an image such that the deep learning model misclassifies the image with a high confidence. Existing adversarial defenses validate their performance using only the classification accuracy. However, classification accuracy by itself is not a reliable metric to determine if the resulting image is "adversarial-free". This is a foundational problem for online image recognition applications where the ground-truth of the incoming image is not known and hence we cannot compute the accuracy of the classifier or validate if the image is "adversarial-free" or not. This paper proposes a novel privacy preserving framework for defending Black box classifiers from adversarial attacks using an ensemble of iterative adversarial image purifiers whose performance is continuously validated in a loop using Bayesian uncertainties. The proposed approach can convert a single-step black box adversarial defense into an iterative defense and proposes three novel privacy preserving Knowledge Distillation (KD) approaches that use prior meta-information from various datasets to mimic the performance of the Black box classifier. Additionally, this paper proves the existence of an optimal distribution for the purified images that can reach a theoretical lower bound, beyond which the image can no longer be purified. Experimental results on six public benchmark datasets namely: 1) Fashion-MNIST, 2) CIFAR-10, 3) GTSRB, 4) MIO-TCD, 5) Tiny-ImageNet, and 6) MS-Celeb show that the proposed approach can consistently detect adversarial examples and purify or reject them against a variety of adversarial attacks.

Analyzing user behavior characteristics in a complex power grid environment is essential for user behavior planning and resource coordination optimization. Traditional user behavior analysis methods based on model-driven and causal analysis have the disadvantages of strong subjectivity and physical models that are difficult to deal with the randomness and uncertainty of user behavior in complex grid environments. In this paper, we use unsupervised learning methods to analyze user behavior in complex power grid environments, and propose user behavior analysis methods based on stacked autoencoder and clustering. We first reduce the complexity of user behavior data by proposing adaptive feature selection algorithm of user behavior based on stacked autoencoder and unsupervised learning (AFS-SAEUL). Finally, we build a user behavior analysis model based on adaptive feature selection and improved clustering (UBA-AFSIC). The model improved the performance of unsupervised classification of user behavior by fusing the adaptive generation strategy of the initial cluster centers. The simulation experiment results on two real electricity datasets and one public electric vehicle charging dataset show that compared with the existing feature selection algorithm and clustering algorithm, the algorithms proposed in this paper have higher feature selection rate and better clustering performance.