Deep neural networks have achieved exceptional performance across a wide range of applications but remain susceptible to adversarial attacks. While most prior research has focused on single-task scenarios, increasing attention is being directed toward adversarial attacks targeting multiple tasks simultaneously. However, existing methods often fail to balance attack performance across tasks in a multi-task model. These approaches typically aim to maximize the model’s overall loss, neglecting task-specific attack difficulties, which results in imbalanced attack performance among tasks. To address this challenge, we propose a novel multi-task adversarial attack method that ensures robust and balanced attack performance across multiple tasks. Our approach dynamically updates task-specific weighting factors through a min-max optimization during the attack, optimizing the worst-case attack performance across all tasks. Experimental results demonstrate that our method significantly enhances the worst-case attack performance across diverse datasets and attack strategies compared to existing approaches. By dynamically adjusting the attack intensity on the least vulnerable tasks, the min-max optimization significantly improves overall attack effectiveness as well as the worst-case performance by balancing the task weights.

Distributed Learning (DL) enables the training of machine learning models across multiple devices, yet it faces challenges like non-IID data distributions and device capability disparities, which can impede training efficiency. Communication bottlenecks further complicate traditional Federated Learning (FL) setups. To mitigate these issues, we introduce the Personalized Federated Learning with Decentralized Selection Training (PFedDST) framework. PFedDST enhances model training by allowing devices to strategically evaluate and select peers based on a comprehensive communication score. This score integrates loss, task similarity, and selection frequency, ensuring optimal peer connections. This selection strategy is tailored to increase local personalization and promote beneficial peer collaborations to strengthen the stability and efficiency of the training process. Our experiments demonstrate that PFedDST not only enhances model accuracy but also accelerates convergence. This approach outperforms state-of-the-art methods in handling data heterogeneity, delivering both faster and more effective training in diverse and decentralized systems.

In this paper, we present a novel hypothesis testing framework to model and analyze an attacker’s decision-making during the reconnaissance phase, and subsequently leverage the framework to characterize optimal deception strategies that can be used to defeat the attacker’s reconnaissance efforts. Our developed model and analytical approaches are well capable of diligently addressing the information-centric nature of the involved attack-defense processes while adapting with the defender’s and attacker’s cognitive biases (irrationalities). Employing the developed hypothesis testing framework, we first characterize a cognitively biased attacker’s optimal prospect theoretic decision rule that enables it to best exploit the information that it acquires during reconnaissance. Leveraging such understanding, we design optimal information falsification strategies that can be employed by a cognitively biased defender to strategically deceive the attacker during its reconnaissance phase under varied considerations. Several numerical results have been presented that provide important insights into the developed strategies.

Federated learning (FL) offers a decentralized approach to machine learning. In FL, models are trained by the data from multiple devices or clients without necessarily centralizing this data, thus preserving privacy and reducing the need for data transfer. Despite its potential, FL faces inherent obstacles, most notably the challenge of achieving a fast convergence rate, especially with large, non-identically distributed client datasets. Also, weight pruning, an effective approach to reduce the number of weight parameters in a deep neural network, is hard to be employed on FL because it involves additional challenges to the convergence between different clients. To deal with the above issues, we propose a unified framework for the convergence and weight pruning in FL. We leverage the inherent structure of the Alternating Direction Method of Multipliers (ADMM) to partition the primary loss function for individual clients and apply specific dual variables to hasten the global model’s convergence. Our method, when tested on MNIST and SVHN datasets, consistently outperforms the established approaches on the convergence rate and model accuracy under the same weight pruning rate. For example, when the ResNet-18 model is pruned by 100 ×, our method achieves 0.65% to 3.09% accuracy improvement for the SVHN dataset under federated learning with non-identically distributed (non-IID) data distribution compared with the established approaches.

Change detection (CD) in heterogeneous remote sensing images has been widely used for disaster monitoring and land-use management. In the past decade, the heterogeneous CD problem has significantly benefited from the development of deep neural networks (DNNs). However, the purely data-driven DNNs perform like a black box where the lack of interpretability limits the trustworthiness and controllability of DNNs in most practical CD applications. As a powerful knowledge-driven tool, copula theory performs well in modeling dependence among random variables. To enhance the interpretability of existing neural networks for heterogeneous CD, we propose a knowledge-data-driven heterogeneous CD method based on a copula-guided neural network, named NN-Copula-CD. In our NN-Copula-CD, the mathematical characteristics of copula are employed as the loss functions to supervise a neural network to learn the dependence between bi-temporal heterogeneous superpixel pairs, and then the changed regions are identified via binary classification based on the degrees of dependence of all the superpixel pairs in the bi-temporal images. We conduct in-depth experiments on four datasets with heterogeneous images, including synthetic aperture radar (SAR), multispectral, and near-infrared images, where quantitative and visual results demonstrate the effectiveness and interpretability of our proposed NN-Copula-CD method.

This paper introduces a representative-based approach for distributed learning that transforms multiple raw data points into a virtual representation. Unlike traditional distributed learning methods such as Federated Learning, which do not offer human interpretability, our method makes complex machine learning processes accessible and comprehensible. It achieves this by condensing extensive datasets into digestible formats, thus fostering intuitive human-machine interactions. Additionally, this approach maintains privacy and communication efficiency, and it matches the training performance of models using raw data. Simulation results show that our approach is competitive with or outperforms traditional Federated Learning in accuracy and convergence, especially in scenarios with complex models and a higher number of clients. This framework marks a step forward in integrating human intuition with machine intelligence, which potentially enhances human-machine learning interfaces and collaborative efforts.

In this work, we consider a binary hypothesis testing problem involving human decision-makers. Due to the nature of human behavior, human decision-makers observe the phenomenon of interest sequentially up to a random length of time. The humans use a belief model to accumulate the log-likelihood ratios until they cease observing the phenomenon. The belief model is used to characterize the perception of the human decision-maker towards observations at different instants of time, i.e., some decision-makers may assign greater importance to observations that were observed earlier, rather than later and vice-versa. We further consider the performance of a group of humans using a global decision-maker that fuses human decisions using the Chair-Varshney rule. When the number of observations that were used by the humans to arrive at their respective decisions are available to the fusion center (FC), the weights in the Chair-Varshney rule are modified to include this information in the decision fusion rule. Numerical and simulation results are presented to corroborate and validate theoretical results.

In distributed detection systems, energy-efficient ordered transmission (EEOT) schemes are able to reduce the number of transmissions required to make a final decision. In this work, we investigate the effect of data falsification attacks on the performance of EEOT-based systems. We derive the probability of error for an EEOT-based system under attack and find an upper bound (UB) on the expected number of transmissions required to make the final decision. Moreover, we tighten this UB by solving an optimization problem via integer programming (IP). We also obtain the FC's optimal threshold which guarantees the optimal detection performance of the EEOT-based system. Numerical and simulation results indicate that it is possible to reduce transmissions while still ensuring the quality of the decision with an appropriately designed threshold.

In critical environments that require a high accuracy of decisions, utilizing human cognitive strengths and expertise in addition to machine observations is advantageous to improve decision quality and enhance situational awareness. While the current literature on human decision making is primarily based on the paradigm of perfect rationality, humans are subject to decision noise and employ stochastic choice rules. Human decision making under such realistic environments needs to be further studied. In this paper, instead of assuming that a human selects the optimal action with probability one, we employ a bounded rationality choice model where all the actions are candidates for selection, but better options are chosen with higher probabilities. In a Bayesian hypothesis testing framework, we evaluate the individual decision making performance when humans have different degrees of bounded rationality. Furthermore, we analyze the decision fusion rule for a team of two human agents and characterize the asymptotic performance of collaborative decision making as the number of human participants becomes large.

While there has been extensive work on modeling of human decision-making both for individuals and groups from a cognitive psychology point of view, research on this topic from a signal processing and information fusion perspective is relatively recent. In this work, we consider a distributed detection problem consisting of a number of human local decision makers and a fusion center (FC). Signal detection theory is exploited to answer why promoting heterogeneity could improve the performance of collaborative human decision-making. We consider the following two scenarios: 1) the local decision makers are independent and the level of heterogeneity is measured in terms of the variability of human expertise and 2) humans make correlated local decisions due to their perceptual and behavioral similarities and heterogeneity is measured by the amount of correlation. In both cases, we show that the detection performance of the FC can be improved with the increase of heterogeneity. In particular, in the second scenario, we develop a portfolio theory-based framework to select participants from correlated human agents so that heterogeneity is enhanced resulting in improved decision-making performance. Simulations are provided for illustration and performance comparison.

Predicting workload using physiological sensors has taken on a diffuse set of methods in recent years. However, the majority of these methods train models on small datasets, with small numbers of channel locations on the brain, limiting a model's ability to transfer across participants, tasks, or experimental sessions. In this paper, we introduce a new method of modeling a large, cross-participant and cross-session set of high density functional near infrared spectroscopy (fNIRS) data by using an approach grounded in cognitive load theory and employing a Bi-Directional Gated Recurrent Unit (BiGRU) incorporating attention mechanism and self-supervised label augmentation (SLA). We show that our proposed CNN-BiGRU-SLA model can learn and classify different levels of working memory load (WML) and visual processing load (VPL) across participants. Importantly, we leverage a multi-label classification scheme, where our models are trained to predict simultaneously occurring levels of WML and VPL. We evaluate our model using leave-one-participant-out (LOOCV) as well as 10-fold cross validation. Using LOOCV, for binary classification (off/on), we reached an F1-score of 0.9179 for WML and 0.8907 for VPL across 22 participants (each participant did 2 sessions). For multi-level (off, low, high) classification, we reached an F1-score of 0.7972 for WML and 0.7968 for VPL. Using 10-fold cross validation, for multi-level classification, we reached an F1-score of 0.7742 for WML and 0.7741 for VPL.

Unlike decision making systems made up of physical sensors where the system parameters are known a priori and can be controlled at will, human behavior in decision making is complex and uncertain. The objective of this work is to study how humans make decisions based on internal and external sources of information under cognitive memory limitations. Due to constrained capacity of working memory, humans are known to perform cognitive tasks and update their beliefs in a sequential manner rather than in parallel. In a Bayesian hypothesis testing framework, we derive the metrics for performance evaluation and comparison when the humans use different ordering of information for processing and to update their beliefs. We show that an appropriate order of information sources can help a cognitive memory limited human make better decisions. Simulations are presented to corroborate the theoretical results.

In this paper, the optimal jamming strategy by an adversary in distributed detection networks is investigated. By utilizing the game-theoretical framework to characterize the interaction between the fusion center (FC) and the jammer as a repeated game, we examine how the behavior of the FC changes with jammer’s strategy. Based on simulation results, we find that the ‘Evenly distributed’ strategy is not always optimal for the jammer. Instead, under certain conditions, the ‘Betting on one channel’ jamming strategy is the best strategy for the jammer.

We study the problem of a sensor performing inference tasks based on the utility theory, where the objective is to derive the optimal resource usage amount that maximizes a profit-cost-based utility function. Furthermore, to enable the concept ofsensing as a servicein the context of IoT systems, we present a market-based paradigm, where there is a “buyer” interested in buying the inference result from the sensor. We jointly optimize the resource usage policy and payment negotiation strategy for the sensor so as to maximize the expected profit. Optimal payment negotiation is analyzed in two situations, namely, when the sensor spends a fixed amount of resource, as well as when the sensor could vary the amount of resource consumption to maximize profit. It is shown that in the presence of the buyer, the optimal amount of resource consumption increases and, hence, the inference accuracy improves. Finally, we present some discussions on how energy efficiency affects the behavior of energy consumption in realistic environments. Simulation results are provided to illustrate the performance of our approach.

In this work, we propose a joint collaboration-compression framework for the random signal detection problem in a resource constrained wireless sensor network (WSN). Specifically, we propose a framework where the local sensors first collaborate (via a linear collaboration matrix) with each other. Then a subset of sensors linearly compress their aggregated information before communicating with the fusion center (FC). We propose a novel metric called generalized deflection coefficient (GDC) for evaluating the detection performance which is shown to be tightly upper bounded by the Kullback-Leibler divergence for Gaussian observations. We jointly design the linear collaboration and compression strategies under power constraints via alternating maximization of the proposed GDC metric. Finally, numerical results are provided to demonstrate the effectiveness of the proposed framework.

Human behavior while decision making is quite complex and uncertain. There are fundamental differences between traditional decision making systems based on sensor data and systems where the agents in the decision making process include humans. The modeling and analysis of human-machine collaborative decision making has become an important research area due to the potential applications in a variety of complex autonomous systems. Incorporating human inputs with physical sensors can be advantageous in enhancing situational assessment for certain situations, and at the same time, brings in technical challenges such as how to characterize the human decision making behavior. In this paper, we discuss some aspects of human-machine networks by focusing on three schemes that include collaborative human decision making with random local thresholds, decision fusion in integrated human-machine networks and binary decision making under cognitive biases. In each case, we aim to optimize the system performance based on appropriate modeling of the human behavior. We also provide a summary of current challenges and research directions related to this problem domain.

This paper considers the design of an auction mechanism for target tracking via crowdsensing. We consider that the crowdsourcing framework consists of a set of sensors, which are embedded in devices belonging to crowd participants, and a fusion center (FC) that uses the quantized measurements from the sensors to track a target. The auction mechanism we develop addresses participatory concerns of the sensors that arise due to energy consumption associated with sensor participation while maximizing the utility of the FC to achieve desired sensing objectives and preventing market manipulations. Moreover, since a crowdsensing environment is typically resource-constrained, in our auction model, we consider that the sensors in the network quantize their private value estimates regarding their energy costs prior to communicating them to the FC. Furthermore, the paper also proposes the concept of selecting a subset of sensors (bidders) to bid (from a set of available sensors) to satisfy resource constraints during the bidding process. Extensive numerical results are provided to gain insights into the proposed mechanism.