Biblio

As machine learning approaches are increasingly used to augment human decision-making, eXplainable Artificial Intelligence (XAI) research has explored methods for communicating system behavior to humans. However, these approaches often fail to account for the affective responses of humans as they interact with explanations. Facial affect analysis, which examines human facial expressions of emotions, is one promising lens for understanding how users engage with explanations. Therefore, in this work, we aim to (1) identify which facial affect features are pronounced when people interact with XAI interfaces, and (2) develop a multitask feature embedding for linking facial affect signals with participants' use of explanations. Our analyses and results show that the occurrence and values of facial AU1 and AU4, and Arousal are heightened when participants fail to use explanations effectively. This suggests that facial affect analysis should be incorporated into XAI to personalize explanations to individuals' interaction styles and to adapt explanations based on the difficulty of the task performed.

The adoption of artificial intelligence (AI) into domains that impact human life (healthcare, agriculture, security and defense, etc.) has led to an increased demand for explainable AI (XAI). Herein, we focus on an under represented piece of the XAI puzzle, information fusion. To date, a number of low-level XAI explanation methods have been proposed for the fuzzy integral (FI). However, these explanations are tailored to experts and its not always clear what to do with the information they return. In this article we review and categorize existing FI work according to recent XAI nomenclature. Second, we identify a set of initial actions that a user can take in response to these low-level statistical, graphical, local, and linguistic XAI explanations. Third, we investigate the design of an interactive user friendly XAI report. Two case studies, one synthetic and one real, show the results of following recommended actions to understand and improve tasks involving classification.

A promising paradigm in protocol design is to hold parties accountable for misbehavior, instead of postulating that they are trustworthy. Recent approaches in defining this property, called accountability, characterized malicious behavior as a deviation from the protocol that causes a violation of the desired security property, but did so under the assumption that all deviating parties are controlled by a single, centralized adversary. In this work, we investigate the setting where multiple parties can deviate with or without coordination in a variant of the applied-π calculus.We first demonstrate that, under realistic assumptions, it is impossible to determine all misbehaving parties; however, we show that accountability can be relaxed to exclude causal dependencies that arise from the behavior of deviating parties, and not from the protocol as specified. We map out the design space for the relaxation, point out protocol classes separating these notions and define conditions under which we can guarantee fairness and completeness. Most importantly, we discover under which circumstances it is correct to consider accountability in the single-adversary setting, where this property can be verified with off-the-shelf protocol verification tools.

Python libraries are widely used for machine learning and scientific computing tasks today. APIs in Python libraries are deprecated due to feature enhancements and bug fixes in the same way as in other languages. These deprecated APIs are discouraged from being used in further software development. Manually detecting and replacing deprecated APIs is a tedious and time-consuming task due to the large number of API calls used in the projects. Moreover, the lack of proper documentation for these deprecated APIs makes the task challenging. To address this challenge, we propose an algorithm and a tool APIScanner that automatically detects deprecated APIs in Python libraries. This algorithm parses the source code of the libraries using abstract syntax tree (ASTs) and identifies the deprecated APIs via decorator, hard-coded warning or comments. APIScanner is a Visual Studio Code Extension that highlights and warns the developer on the use of deprecated API elements while writing the source code. The tool can help developers to avoid using deprecated API elements without the execution of code. We tested our algorithm and tool on six popular Python libraries, which detected 838 of 871 deprecated API elements. Demo of APIScanner: https://youtu.be/1hy\_ugf-iek. Documentation, tool, and source code can be found here: https://rishitha957.github.io/APIScanner.

Smart grid achieves reliable, efficient and flexible grid data processing by integrating traditional power grid with information and communication technology. The control center can evaluate the supply and demand of the power grid through aggregated data of users, and then dynamically adjust the power supply, price of the power, etc. However, since the grid data collected from users may disclose the user's electricity using habits and daily activities, the privacy concern has become a critical issue. Most of the existing privacy-preserving data collection schemes for smart grid adopt homomorphic encryption or randomization techniques which are either impractical because of the high computation overhead or unrealistic for requiring the trusted third party. In this paper, we propose a privacy-preserving smart grid data aggregation scheme satisfying local differential privacy (LDP) based on randomized response. Our scheme can achieve efficient and practical estimation of the statistics of power supply and demand while preserving any individual participant's privacy. The performance analysis shows that our scheme is efficient in terms of computation and communication overhead.

Researchers have demonstrated the effectiveness of self-explanations (SE) as an instructional practice and study strategy. However, there is a lack of work studying the characteristics of SE responses prompted by collaborative activities. In this paper, we use writing analytics to investigate differences between SE text responses resulting from individual versus collaborative learning activities. A Coh-Metrix analysis suggests that students in the collaborative SE activity demonstrated a higher level of comprehension. Future research should explore how writing analytics can be incorporated into CSCL systems to support student performance of SE activities.

For the task of natural language processing (NLP), Word embedding technology has a certain impact on the accuracy of deep neural network algorithms. Considering that the current word embedding method cannot realize the coexistence of words and phrases in the same vector space. Therefore, we propose an enhanced word embedding (EWE) method. Before completing the word embedding, this method introduces a unique sentence reorganization technology to rewrite all the sentences in the original training corpus. Then, all the original corpus and the reorganized corpus are merged together as the training corpus of the distributed word embedding model, so as to realize the coexistence problem of words and phrases in the same vector space. We carried out experiment to demonstrate the effectiveness of the EWE algorithm on three classic benchmark datasets. The results show that the EWE method can significantly improve the classification performance of the CNN model.

Nowadays, the Internet of Things (IoT) is playing an important role in our life. This inevitably generates mass data and requires a more secure transmission. As blockchain technology can build trust in a distributed environment and ensure the data traceability and tamper resistance, it is a promising way to support IoT data transmission and sharing. In this paper, edge computing is considered to provide adequate resources for end users to offload computing tasks in the blockchain enabled IoT system, and the node pairing problem between end users and edge computing servers is researched with the consideration of wireless channel quality and the service quality. From the perspective of the end users, the objective optimization is designed to maximize the profits and minimize the payments for completing the tasks and ensuring the resource limits of the edge servers at the same time. The deep reinforcement learning (DRL) method is utilized to train an intelligent strategy, and the policy gradient based node pairing (PG-NP) algorithm is proposed. Through a deep neural network, the well-trained policy matched the system states to the optimal actions. The REINFORCE algorithm with baseline is applied to train the policy network. According to the training results, as the comparison strategies are max-credit, max-SINR, random and max-resource, the PG-NP algorithm performs about 57% better than the second-best method. And testing results show that PGNP also has a good generalization ability which is negatively correlated with the training performance to a certain extend.

We study the problem of generation of cycles, chains and graphs of pairing-friendly elliptic curves using in succinct non-interactive arguments for knowledge protocols in blockchain. The task to build a “stick” for existing MNT753 cycle is reduced to the factorization problem for big numbers. Together with graphs of pairing friendly elliptic curves we consider auxiliary graphs of their orders (primes or irreducible polynomials) associated to vertices and embedding degrees to edges. Numerical experiments allow us to conjecture that (except of MNT case): 1) for any fixed embedding degrees there exist only finite number of such cycles and, hence, there are no families of such cycles; 2) chains of prime order are very rare; we suppose that there are no polynomial families of such chains. It is hard to find a family of pairing friendly elliptic curves with the base field order q(x) such that ζk ∈ Q[x]/(q(x)) for k \textbackslashtextgreater 6. From other hand our examples show that we can apply Brezing-Weng construction with k=6 and D=3 iteratively to obtain chains of length 3-4. We build 1) a family of 1-chains with embedding degrees 8 and 7, where all orders are given by cyclotomic polynomials; 2) a combination of MNT cycle and near-MNT curve.

The design of large-scale quantum networks and any future quantum internet will rely on quantum repeaters and how quantum information flows through it. Tasks performed on such networks will go well beyond quantum key distribution and are likely to include quantum remote sensing and distributed quantum computation. In this presentation we will explore the various ways that such networks could be designed to support those advanced tasks. Critical to this will be quantum routing which we should is highly dependent on the repeater architecture. We introduce a quantum quality of service to help characterize the systems performance and shows how it leads interesting network and routing behavior.

Dynamic taint analysis (DTA) is widely used by various applications to track information flow during runtime execution. Existing DTA techniques use rule-based taint-propagation, which is neither accurate (i.e., high false positive rate) nor efficient (i.e., large runtime overhead). It is hard to specify taint rules for each operation while covering all corner cases correctly. Moreover, the overtaint and undertaint errors can accumulate during the propagation of taint information across multiple operations. Finally, rule-based propagation requires each operation to be inspected before applying the appropriate rules resulting in prohibitive performance overhead on large real-world applications.In this work, we propose Neutaint, a novel end-to-end approach to track information flow using neural program embeddings. The neural program embeddings model the target's programs computations taking place between taint sources and sinks, which automatically learns the information flow by observing a diverse set of execution traces. To perform lightweight and precise information flow analysis, we utilize saliency maps to reason about most influential sources for different sinks. Neutaint constructs two saliency maps, a popular machine learning approach to influence analysis, to summarize both coarse-grained and fine-grained information flow in the neural program embeddings.We compare Neutaint with 3 state-of-the-art dynamic taint analysis tools. The evaluation results show that Neutaint can achieve 68% accuracy, on average, which is 10% improvement while reducing 40× runtime overhead over the second-best taint tool Libdft on 6 real world programs. Neutaint also achieves 61% more edge coverage when used for taint-guided fuzzing indicating the effectiveness of the identified influential bytes. We also evaluate Neutaint's ability to detect real world software attacks. The results show that Neutaint can successfully detect different types of vulnerabilities including buffer/heap/integer overflows, division by zero, etc. Lastly, Neutaint can detect 98.7% of total flows, the highest among all taint analysis tools.

Energy security program which becomes the part of energy management must ensure the high reliability of the electric power transmission system so that the customer can be served very well. However, there are several problems that can hinder reliability achievement such as the long duration of fault recovery. On the other side, the prediction of fault recovery duration becomes a very challenging task. Because there are still few machine learning-based solution offer this paper proposes a machine learning-based strategy by using Naive-Bayes Classifier (NBC) and Support Vector Machine (SVM) in predicting the fault recovery duration class. The dataset contains 3398 rows of non-temporary-fault type records, six input features (Substation, Asset Type, Fault Category, Outage Start Time, Outage Day, and Outage Month) and single target feature (Fault Recovery Duration). According to the performance test result, those two methods reach around 97-99% of accuracy, average sensitivity, and average specificity. In addition, one of the advantages obtained in field of fault recovery prediction is increasing the accuracy of likelihood level calculation of the long fault recovery time risk.

The Internet of Things (IoT) is an evolving network of billions of interconnected physical objects, such as, numerous sensors, smartphones, wearables, and embedded devices. These physical objects, generally referred to as the smart objects, when deployed in real-world aggregates useful information from their surrounding environment. As-of-late, this notion of IoT has been extended to incorporate the social networking facets which have led to the promising paradigm of the `Social Internet of Things' (SIoT). In SIoT, the devices operate as an autonomous agent and provide an exchange of information and services discovery in an intelligent manner by establishing social relationships among them with respect to their owners. Trust plays an important role in establishing trustworthy relationships among the physical objects and reduces probable risks in the decision making process. In this paper, a trust computational model is proposed to extract individual trust features in a SIoT environment. Furthermore, a machine learning-based heuristic is used to aggregate all the trust features in order to ascertain an aggregate trust score. Simulation results illustrate that the proposed trust-based model isolates the trustworthy and untrustworthy nodes within the network in an efficient manner.

The exceptional attributes of impromptu network of being framework less, self-composed and unconstrained make the task more challenging to secure it. In mobile Ad-hoc network nodes reliant on one another for transmitting information, that make MANET helpless against different sorts of security attacks. These security attacks can be arranged as Passive and Active attacks. Wormhole is an Active attack and considered generally risky as it can make significant harm routing. Various secure routing mechanism has been created are based on cryptography mechanism, need pre-organized structure, centralized authority, or need external hardware, etc. These components are unreasonable due to restricted accessible assets in MANET. In this paper, we are proposing an effective trust-based mechanism based on the concept of Node to Node packet delay for the detection of the malevolent node of wormhole. The trust value of each node is calculated by observing the packet transaction among adjacent nodes and later this trust value is used for identification of malevolent node. Based on the trust values, further routing decisions and selecting a secured route can be perform.

RPL, the IPv6 Routing Protocol for low-power and lossy networks, was standardized by the Internet Engineering Task Force (IETF) in 2011. It is developed to connect resource constrained devices enabled by low-power and lossy networks (LLNs). RPL prominently becomes the routing protocol for IoT. However, the RPL protocol is facing many challenges such as trustworthiness among the nodes which need to be addressed and resolved to make the network secure and efficient. In this paper, a multicasting technique is developed that is based on trust mechanism to resolve this issue. This mechanism manages and protects the network from untrusted nodes which can hamper the security and result in delayed and distorted transmission of data. It allows any node to decide whether to trust other nodes or not during the construction of the topology. This is then proved efficient by comparing it with broadcasting nature of the transmission among the nodes in terms of energy, throughput, percentage of alive and dead nodes.

Visual Analytics uses data visualization techniques for enabling compelling data analysis by engaging graphical and visual portrayal. In the domain of cybersecurity, convincing visual representation of data enables to ascertain valuable observations that allow the domain experts to construct efficient cyberattack mitigation strategies and provide useful decision support. We present a survey of visual analytics tools and methods in the domain of cybersecurity. We explore and discuss Technical Threat Intelligence visualization tools using the Five Question Method. We conclude the analysis of the works using Moody's Physics of Notations, and VIS4ML ontology as a methodological background of visual analytics process. We summarize our analysis as a high-level model of visual analytics for cybersecurity threat analysis.

Despite the performance of state-of-the-art Artificial Intelligence (AI) systems, some sectors hesitate to adopt AI because of a lack of trust in these systems. This attitude is prevalent among high-risk areas, where there is a reluctance to remove humans entirely from the loop. In these scenarios, Augmentation provides a preferred alternative over complete Automation. Instead of replacing humans, AI Augmentation uses AI to improve and support human operations, creating an environment where humans work side by side with AI systems. In this paper, we discuss how AI Augmentation can provide a path for building Trustworthy AI. We exemplify this approach using Robot Teleoperation. We lay out design guidelines and motivations for the development of AI Augmentation for Robot Teleoperation. Finally, we discuss the design of a Robot Teleoperation testbed for the development of AI Augmentation systems.

Endpoint Detection and Response (EDR) tools provide visibility into sophisticated intrusions by matching system events against known adversarial behaviors. However, current solutions suffer from three challenges: 1) EDR tools generate a high volume of false alarms, creating backlogs of investigation tasks for analysts; 2) determining the veracity of these threat alerts requires tedious manual labor due to the overwhelming amount of low-level system logs, creating a "needle-in-a-haystack" problem; and 3) due to the tremendous resource burden of log retention, in practice the system logs describing long-lived attack campaigns are often deleted before an investigation is ever initiated.This paper describes an effort to bring the benefits of data provenance to commercial EDR tools. We introduce the notion of Tactical Provenance Graphs (TPGs) that, rather than encoding low-level system event dependencies, reason about causal dependencies between EDR-generated threat alerts. TPGs provide compact visualization of multi-stage attacks to analysts, accelerating investigation. To address EDR's false alarm problem, we introduce a threat scoring methodology that assesses risk based on the temporal ordering between individual threat alerts present in the TPG. In contrast to the retention of unwieldy system logs, we maintain a minimally-sufficient skeleton graph that can provide linkability between existing and future threat alerts. We evaluate our system, RapSheet, using the Symantec EDR tool in an enterprise environment. Results show that our approach can rank truly malicious TPGs higher than false alarm TPGs. Moreover, our skeleton graph reduces the long-term burden of log retention by up to 87%.

The discovery of adversarial examples poses a serious risk to the deep neural networks (DNN). By adding a subtle perturbation that is imperceptible to the human eye, a well-behaved DNN model can be easily fooled and completely change the prediction categories of the input samples. However, research on adversarial attacks in the field of modulation recognition mainly focuses on increasing the prediction error of the classifier, while ignores the importance of decreasing the perceptual invisibility of attack. Aiming at the task of DNNbased modulation recognition, this study designs the Fitting Difference as a metric to measure the perturbed waveforms and proposes a new method: the Nesterov Adam Iterative Method to generate adversarial examples. We show that the proposed algorithm not only exerts excellent white-box attacks but also can initiate attacks on a black-box model. Moreover, our method decreases the waveform perceptual invisibility of attacks to a certain degree, thereby reducing the risk of an attack being detected.

This paper presents an introduction to functional safety through ISO 26262 focusing on system, software and hardware possible failures that bring security threats and discussion on DO 254. It discusses the approach to bridge the gap between different other hazard level and system ability to identify the particular fault and resolve it minimum time span possible. Results are analyzed by designing models to check and avoid all the failures, loophole prior development.

In recent years, several trust management frame-works and models have been proposed for the Internet of Things (IoT). Focusing primarily on distributed trust management schemes; testing and validation of these models is still a challenging task. It requires the implementation of the proposed trust model for verification and validation of expected outcomes. Nevertheless, a stand-alone and standard IoT network simulator for testing of distributed trust management scheme is not yet available. In this paper, a .NET-based Distributed Trust Management Simulator for IoT Networks (DTMSim-IoT) is presented which enables the researcher to implement any static/dynamic trust management model to compute the trust value of a node. The trust computation will be calculated based on the direct-observation and trust value is updated after every transaction. Transaction history and logs of each event are maintained which can be viewed and exported as .csv file for future use. In addition to that, the simulator can also draw a graph based on the .csv file. Moreover, the simulator also offers to incorporate the feature of identification and mitigation of the On-Off Attack (OOA) in the IoT domain. Furthermore, after identifying any malicious activity by any node in the networks, the malevolent node is added to the malicious list and disseminated in the network to prevent potential On-Off attacks.

The privacy protection and information security are two crucial issues for future advanced artificial intelligence devices, especially for cooperative system with rich core data exchange which may offer opportunities for attackers to fake interaction messages. To combat such threat, great efforts have been made by introducing trust mechanism in initiative or passive way. Furthermore, blockchain and distributed ledger technology provide a decentralized and peer-to-peer network, which has great potential application for multi-agent system, such as IoTs and robots. It eliminates third-party interference and data in the blockchain are stored in an encrypted way permanently and anti-destroys. In this paper, a methodology of blockchain is proposed and designed for advanced cooperative system with artificial intelligence to protect privacy and sensitive data exchange between multi-agents. The validation procedure is performed in laboratory by a three-level computing networks of Raspberry Pi 3B+, NVIDIA Jetson Tx2 and local computing server for a robot system with four manipulators and four binocular cameras in peer computing nodes by Go language.

Software Defined Networking (SDN) is disruptive networking technology which adopts a centralised framework to facilitate fine-grained network management. However security in SDN is still in its infancy and there is need for significant work to deal with different attacks in SDN. In this paper we discuss some of the possible attacks on SDN switches and propose techniques for detecting the attacks on switches. We have developed a Switch Security Application (SSA)for SDN Controller which makes use of trusted computing technology and some additional components for detecting attacks on the switches. In particular TPM attestation is used to ensure that switches are in trusted state during boot time before configuring the flow rules on the switches. The additional components are used for storing and validating messages related to the flow rule configuration of the switches. The stored information is used for generating a trusted report on the expected flow rules in the switches and using this information for validating the flow rules that are actually enforced in the switches. If there is any variation to flow rules that are enforced in the switches compared to the expected flow rules by the SSA, then, the switch is considered to be under attack and an alert is raised to the SDN Administrator. The administrator can isolate the switch from network or make use of trusted report for restoring the flow rules in the switches. We will also present a prototype implementation of our technique.

Knowledge Tracing (KT) is a task that aims to assess students' mastery level of knowledge and predict their performance over questions, which has attracted widespread attention over the years. Recently, an increasing number of researches have applied deep learning techniques to knowledge tracing and have made a huge success over traditional Bayesian Knowledge Tracing methods. Most existing deep learning-based methods utilized either Recurrent Neural Networks (RNNs) or Convolutional Neural Networks (CNNs). However, it is worth noticing that these two sorts of models are complementary in modeling abilities. Thus, in this paper, we propose a novel knowledge tracing model by taking advantage of both two models via combining them into a single integrated model, named Convolutional Recurrent Knowledge Tracing (CRKT). Extensive experiments show that our model outperforms the state-of-the-art models in multiple KT datasets.

One of the most significant systems in computer network security assurance is the assessment of computer network security. With the goal of finding an effective method for performing the process of security evaluation in a computer network, this paper uses a deep neural network to be responsible for the task of security evaluating. The DNN will be built with python on Spyder IDE, it will be trained and tested by 17 network security indicators then the output that we get represents one of the security levels that have been already defined. The maj or purpose is to enhance the ability to determine the security level of a computer network accurately based on its selected security indicators. The method that we intend to use in this paper in order to evaluate network security is simple, reduces the human factors interferences, and can obtain the correct results of the evaluation rapidly. We will analyze the results to decide if this method will enhance the process of evaluating the security of the network in terms of accuracy.