Radio frequency fingerprint identification (RFFI) aims to identify subtle impairments in hardware devices, which play an important role in the mobile environment security community. To identify various mobile devices in the complex electromagnetic environment, deep learning methods such as convolutional neural networks (CNN) and recurrent neural networks (RNN) have been adopted to extract device hardware-related features. However, the single network structure has difficulty in comprehensive feature extraction, as many factors can introduce hardware impairments. In this paper, we propose a hybrid model termed switching dynamical deep network (SDDN) for RFFI tasks, which can jointly extract both coarse-grained radio frequency fingerprints (RFFs) and fine-grained RFFs. Additionally, the proposed hybrid model consists of a probabilistic part and a deterministic part. Specifically, in the probabilistic part, the switching linear dynamical systems (SLDS) are incorporated to establish the correspondence between the signal slice and the feature extraction network (FEN). In the deterministic part, multiple independent FENs are established to extract the RFFs. Moreover, to automatically determine the suitable number of FENs, a Bayesian nonparametric prior distribution is placed over the probabilistic part. Finally, an end-to-end parameter optimization method that is based on variational inference and stochastic gradient descent is proposed. Experiments on a real-life Wi-Fi dataset demonstrate the superiority of the proposed method over existing methods.

Due to the remarkable reasoning ability, Large language models (LLMs) have demonstrated impressive performance in knowledge graph question answering (KGQA) tasks, which find answers to natural language questions over knowledge graphs (KGs). To alleviate the hallucinations and lack of knowledge issues of LLMs, existing methods often retrieve the question-related information from KGs to enrich the input context. However, most methods focus on retrieving the relevant information while ignoring the importance of different types of knowledge in reasoning, which degrades their performance. To this end, this paper reformulates the KGQA problem as a graphical model and proposes a three-stage framework named the Evidence Path Enhanced Reasoning Model (EPERM) for KGQA. In the first stage, EPERM uses the fine-tuned LLM to retrieve a subgraph related to the question from the original knowledge graph. In the second stage, EPERM filters out the evidence paths that faithfully support the reasoning of the questions, and score their importance in reasoning. Finally, EPERM uses the weighted evidence paths to reason the final answer. Since considering the importance of different structural information in KGs for reasoning, EPERM can improve the reasoning ability of LLMs in KGQA tasks. Extensive experiments on benchmark datasets demonstrate that EPERM achieves superior performances in KGQA tasks.

Domain generalization aims to learn a model from multiple training domains and generalize it to unseen test domains. Recent theory has shown that seeking the deep models, whose parameters lie in the flat minima of the loss landscape, can significantly reduce the out-of-domain generalization error. However, existing methods often neglect the consistency of loss landscapes in different domains, resulting in models that are not simultaneously in the optimal flat minima in all domains, which limits their generalization ability. To address this issue, this paper proposes an iterative Self-Feedback Training (SFT) framework to seek consistent flat minima that are shared across different domains by progressively refining loss landscapes during training. It alternatively generates a feedback signal by measuring the inconsistency of loss landscapes in different domains and refines these loss landscapes for greater consistency using this feedback signal. Benefiting from the consistency of the flat minima within these refined loss landscapes, our SFT helps achieve better out-of-domain generalization. Extensive experiments on DomainBed demonstrate superior performances of SFT when compared to state-of-the-art sharpness-aware methods and other prevalent DG baselines. On average across five DG benchmarks, SFT surpasses the sharpness-aware minimization by 2.6% with ResNet-50 and 1.5% with ViT-B/16, respectively.

Automatic modulation classification (AMC) plays a pivotal role in radar and communication systems. Traditional AMC methods assume an equal number of samples for each modulation during training. However, in real-world scenarios, the number of samples collected for different modulations can vary significantly. This disparity leads the model to overfit to majority classes while underrepresenting minority classes in the optimization process, ultimately leading to degraded classification performance. To tackle this issue, this paper presents a Self-supervised Aligned Data Augmentation Network (SADA-Net) for imbalanced AMC. We leverage Data Augmentation (DA) not only to balance data distribution but also increase data diversity, boosting the model’s robustness in handling minority classes. To amplify the effectiveness of DA, a self-supervised aligned module is incorporated to maintain semantic consistency, preventing the model from focusing on irrelevant variations. Furthermore, an adaptive fusion learning strategy is proposed to dynamically adjust the focus between majority classes and minority classes during training process. This progressive training strategy avoids damaging the learned universal features from majority classes when emphasizing the minority data, ensuring a balanced feature learning process. Comprehensive experiments on simulated, publicly available and real-world datasets demonstrate the effectiveness and generalization of SADA-Net under various class-imbalanced conditions.

Driven by the pursuit of enhanced performance, deep learning has recently seen rapid developments in the scaling of network architectures and parameters. However, this advancement has led to extremely high computational costs, undesirable in real-time and resource-limited scenarios. To address these challenges, we propose SwiftNet, a cost-efficient deep learning framework. Our novelty lies in SwiftNet's innovative multidimensional early-exit strategy that integrates seamlessly with existing neural network architectures. The framework includes additional branch classifiers concatenated to the backbone network, allowing high-confidence samples to exit early, thereby, reducing computational load. Unlike traditional methods, SwiftNet dynamically assesses confidence levels, ensuring only low-confidence samples proceed to subsequent classifiers or the final layer, optimizing resource usage without compromising accuracy. We have validated SwiftNet on multiple neural network models and datasets, demonstrating its ability to significantly reduce the computational cost of models while maintaining neural network performance.

Policy space response oracles (PSRO) is a promising tool to find an approximate Nash equilibrium (NE) in a two-player zero-sum game. It solves the equilibrium by iteratively expanding a small-scale meta-game formed by a restricted strategy population consisting of historical approximate best responses of the meta-games. However, since these best responses have a strong correlation with each other, existing PSRO and its variants often have the slow diversity growth of the strategy population, and thus suffer from poor exploration efficiency and slow convergence rate. To address this problem, this article proposes Purified PSRO, which deliberately maintains a pure strategy population formed by pure strategy bases of approximate best responses. A novel module namely non-best response suppression (NBRS) is introduced to calculate a pure strategy base with better orthogonality to expand the strategy population at each epoch. In this way, Purified PSRO can quickly increase the diversity of the strategy population, thus greatly enhance the efficiency of exploration. Theoretically, we prove the convergence of Purified PSRO. Moreover, we introduce an early stop module to reduce computation cost, and give the upper bound of the exploitability when the algorithm stops early. Extensive experiments on random games of skill (RGoS) and real-world meta-games show that Purified PSRO can consistently outperform existing SOTA methods, sometimes with a large margin.

This article studies how to learn approximate Nash equilibrium (NE) from static historical datasets by empirical game-theoretic analysis (EGTA), which provides a simulation-based framework to model complex multiagent interactions. Generally, EGTA requires plentiful interactions with the environment or simulator to estimate a cogent and tractable game model approximating the underlying game. However, these exploratory interactions often suffer from low data utilization efficiency and may not be feasible in risk-sensitive applications. To address these problems, this article investigates a new EGTA paradigm for offline settings and introduces a novel algorithm called conservative offline policy space response oracle (COPSRO) to identify NE from fixed datasets without active data collection. COPSRO initiates by extracting a set of strategies from the offline dataset to construct an overcomplete strategy population, achieving an approximation to the policy space of the original game. Then, COPSRO integrates the conservative critic (CC) to tackle the challenge of overestimation inherent in offline learning scenarios. Additionally, it devises the offline NE solver to iteratively compute approximate NE. Consequently, COPSRO can ascertain equilibrium strategies without real-world interaction, markedly enhancing its utility in risk-averse settings. This article provides both theoretical analysis and empirical evaluation to demonstrate the effectiveness and superiority of COPSRO across various real-world tasks in the offline setting. Our method surpasses existing approaches in terms of convergence and exploitability, especially when the coverage ration of dataset is low (20% or 10%).

Automatic modulation recognition plays a critical role in modern communication systems as it provides valuable information for tasks such as signal recognition, interference jamming, and interception. Despite the significant potential of machine learning in this field, its practical effectiveness often falls short of expectations. In non-cooperative communication systems, irregular phase shifts and carrier frequency offsets (CFO) in the signal can negatively impact recognition accuracy. Conventional algorithms for compensating CFO, including those reliant on phase-locked loops, as well as various existing non-informative modulation techniques, encounter constraints in non-cooperative environments. This paper presents a novel CFO self-compensating neural network. A CFO compensation module is designed and integrated into the neural network. Additionally, a regularization term is proposed to evaluate the effectiveness of CFO compensation, which is incorporated into the loss function. The effectiveness of the proposed approach is demonstrated through experiments conducted on both publicly accessible datasets and experimentally measured data generated using software-defined radio. Experimental results demonstrate that the proposed new architecture maintains reliable recognition capabilities even in the presence of CFO. Additionally, it can be applied to all modulation schemes that are demodulable using constellation diagrams in non-cooperative conditions.

Knowledge graph embedding (KGE) is an efficient and scalable method for knowledge graph completion. However, most existing KGE methods suffer from the challenge of multiple relation semantics, which often degrades their performance. This is because most KGE methods learn fixed continuous vectors for entities (relations) and make deterministic entity predictions to complete the knowledge graph, which hardly captures multiple relation semantics. To tackle this issue, previous works try to learn complex probabilistic embeddings instead of fixed embeddings but suffer from heavy computational complexity. In contrast, this paper proposes a simple yet efficient framework namely the Knowledge Graph Diffusion Model (KGDM) to capture the multiple relation semantics in prediction. Its key idea is to cast the problem of entity prediction into conditional entity generation. Specifically, KGDM estimates the probabilistic distribution of target entities in prediction through Denoising Diffusion Probabilistic Models (DDPM). To bridge the gap between continuous diffusion models and discrete KGs, two learnable embedding functions are defined to map entities and relation to continuous vectors. To consider connectivity patterns of KGs, a Conditional Entity Denoiser model is introduced to generate target entities conditioned on given entities and relations. Extensive experiments demonstrate that KGDM significantly outperforms existing state-of-the-art methods in three benchmark datasets.

Knowledge graph embedding (KGE) is an efficient and scalable method for knowledge graph completion tasks. Existing KGE models typically map entities and relations into a unified continuous vector space and define a score function to capture the connectivity patterns among the elements (entities and relations) of facts. The score on a fact measures its plausibility in a knowledge graph (KG). However, since the connectivity patterns are very complex in a real knowledge graph, it is difficult to define an explicit and efficient score function to capture them, which also limits their performance. This paper argues that plausible facts in a knowledge graph come from a distribution in the low-dimensional fact space. Inspired by this insight, this paper proposes a novel framework called Fact Embedding through Diffusion Model (FDM) to address the knowledge graph completion task. Instead of defining a score function to measure the plausibility of facts in a knowledge graph, this framework directly learns the distribution of plausible facts from the known knowledge graph and casts the entity prediction task into the conditional fact generation task. Specifically, we concatenate the elements embedding in a fact as a whole and take it as input. Then, we introduce a Conditional Fact Denoiser to learn the reverse denoising diffusion process and generate the target fact embedding from noised data. Extensive experiments demonstrate that FDM significantly outperforms existing state-of-the-art methods in three benchmark datasets.

While the Internet of Things (IoT) technology is booming and offers huge opportunities for information exchange, it also faces unprecedented security challenges. As an important complement to the physical-layer security technologies for IoT, radio frequency fingerprinting (RFF) is of great interest due to its difficulty in counterfeiting. Recently, many machine learning (ML)-based RFF algorithms have emerged. In particular, deep learning (DL) has shown great benefits in automatically extracting complex and subtle features from raw data with high-classification accuracy. However, DL algorithms face the computational cost problem as the difficulty of the RFF task and the size of the deep neural network have increased dramatically. To address the above challenge, this article proposes a novel cost-effective early-exit neural network consisting of a complex-valued neural network (CVNN) backbone with multiple random forest branches, called hybrid CVNN-RF. Unlike conventional studies that use a single fixed DL model to process all radio frequency (RF) samples, our hybrid CVNN-RF considers differences in the recognition difficulty of RF samples and introduces an early-exit mechanism to dynamically process the samples. When processing “easy” samples that can be well classified with high confidence, the hybrid CVNN-RF can end early at the random forest branch to reduce computational cost. Conversely, subsequent network layers will be activated to ensure accuracy. To further improve the early-exit rate, an automated multidimensional early-exit strategy is proposed to achieve scheduling control from multiple dimensions within the network depth and classification category. Finally, our experiments on the public ADS-B data set show that the proposed algorithm can reduce the computational cost by 83% while improving the accuracy by 1.6% under a classification task with 100 categories.

Radio Frequency Fingerprint Identification (RFFI), which exploits non-ideal hardware-induced unique distortion resident in the transmit signals to identify an emitter, is emerging as a means to enhance the security of communication systems. Recently, machine learning has achieved great success in developing state-of-the-art RFFI models. However, few works consider cross-receiver RFFI problems, where the RFFI model is trained and deployed on different receivers. Due to altered receiver characteristics, direct deployment of RFFI model on a new receiver leads to significant performance degradation. To address this issue, we formulate the cross-receiver RFFI as a model adaptation problem, which adapts the trained model to unlabeled signals from a new receiver. We first develop a theoretical generalization error bound for the adaptation model. Motivated by the bound, we propose a novel method to solve the cross-receiver RFFI problem, which includes domain alignment and adaptive pseudo-labeling. The former aims at finding a feature space where both domains exhibit similar distributions, effectively reducing the domain discrepancy. Meanwhile, the latter employs a dynamic pseudo-labeling scheme to implicitly transfer the label information from the labeled receiver to the new receiver. Experimental results indicate that the proposed method can effectively mitigate the receiver impact and improve the cross-receiver RFFI performance.

Open-set Temporal Action Localization (OSTAL) is a critical and challenging task that aims to recognize and temporally localize human actions in untrimmed videos in open word scenarios. The main challenge in this task is the knowledge transfer from known actions to unknown actions. However, existing methods utilize limited training data and overparameterized deep neural network, which have poor generalization. This paper proposes a novel Generalized OSTAL model (namely GOTAL) to learn generalized representations of actions. GOTAL utilizes a Transformer network to model actions and a open-set detection head to perform action localization and recognition. Benefitting from Transformer's temporal modeling capabilities, GOTAL facilitates the extraction of human motion information from videos to mitigate the effects of irrelevant background data. Furthermore, a sharpness minimization algorithm is used to learn the network parameters of GOTAL, which facilitates the convergence of network parameters towards flatter minima by simultaneously minimizing the training loss value and sharpness of the loss plane. The collaboration of the above components significantly enhances the generalization of the representation. Experimental results demonstrate that GOTAL achieves the state-of-the-art performance on THUMOS14 and ActivityNet1.3 benchmarks, confirming the effectiveness of our proposed method.

Recently, various wireless Internet-of-Things devices and unmanned aerial vehicles have been frequently used to facilitate daily life. On the other hand, they can also pose serious threats to public security if maliciously used. A countermeasure against these threats is to transmit jamming signals to break the communication links between attacker and wireless devices. However, this is not easy since modern multi-device (multi-user) wireless communication systems are smart in avoiding jamming. Moreover, a relay is widely used to improve the jamming resistance. To successfully jam those malicious devices, we consider a jamming and anti-jamming zero-sum game, in which the attacker tries to maximize the sum-rate of a relay-assisted wireless system and the jammer tries to minimize it. Finding the equilibrium is challenging due to its non-convexity and complex structure. We address this problem in two cases. Specifically, in the single-device (single-user) case, we show that putting the whole jamming power to either relay or device achieves the Nash equilibrium. In the multi-device case, an efficient method is developed, which transforms this game into a min-max optimization problem and decouples it into two sub-problems by the hybrid block successive approximation method. The resultant sub-problems are solved by the multi-block alternating direction method of multipliers and the block successive upper-bound minimization method of multipliers, respectively. Then, a stationary point of the original problem can be iteratively achieved. Numerical results demonstrate that the two proposed jamming methods outperform many traditional methods.

Query embedding (QE)-which aims to embed entities and first-order logical (FOL) queries in a vector space, has shown great power in answering FOL queries on knowledge graphs (KGs).Existing QE methods divide a complex query into a sequence of mini-queries according to its computation graph and perform logical operations on the answer sets of mini-queries to get answers.However, most of them assume that answer sets satisfy an individual distribution (e.g., Uniform, Beta, or Gaussian), which is often violated in real applications and limit their performance.In this paper, we propose a Neural-based Mixture Probabilistic Query Embedding Model (NMP-QEM) that encodes the answer set of each mini-query as a mixed Gaussian distribution with multiple means and covariance parameters, which can approximate any random distribution arbitrarily well in real KGs.Additionally, to overcome the difficulty in defining the closed solution of negation operation, we introduce neural-based logical operators of projection, intersection and negation for a mixed Gaussian distribution to answer all the FOL queries.Extensive experiments demonstrate that NMP-QEM significantly outperforms existing stateof-the-art methods on benchmark datasets.In NELL995, NMP-QEM achieves a 31% relative improvement over the state-of-the-art.

Low probability of intercept (LPI) radar signals have been widely used in modern radars due to the advantages of being hardly intercepted by non-cooperative receivers. Therefore, the waveform recognition of LPI radar signals has recently gained increasing attention. In this letter, an improved LPI radar waveform recognition framework with local dense connection U-net and novel loss function is proposed. Specifically, radar signals are transformed into time-frequency images (TFIs) through time-frequency analysis techniques. Then, the local dense connection U-net is proposed to reduce interference of noise and enhance TFIs features. Finally, the recognition task is implemented through a deep convolutional neural network, which is trained with a noise-robust loss function called SSR-loss. Compared with other existing LPI radar waveform recognition frameworks, our method can obtain a significant improvement in recognition accuracy under noise conditions. When signal-to-noise ratio (SNR) is as low as −10 dB, the overall probability of successful recognition on twelve kinds of typical modulated signals can reach up to 91.17%.

Knowledge graphs (KGs) often encounter knowledge incompleteness, necessitating a demand for KG completion. Path-based methods are one of the most important approaches to this task. However, since the number of entities is much larger than that of relations in a knowledge graph, existing path-based methods are only used to predict the relations between entity pairs, and are rarely applied to solve the entity prediction task. To address the issue, this paper proposes a new framework called Path Ranking Model (PRM) for the knowledge graph completion task. Our key idea is to exploit both the observable patterns and latent semantic information in relation paths to predict the entities. Extensive experiments on public popular datasets demonstrate the effectiveness of our proposed framework in the entity prediction task.

In an ever-increasingly complex electromagnetic environment, multiple low probability of intercept (LPI) radar emitters may transmit their own signals simultaneously on similar bands, resulting in overlapping receiving signals in both time and frequency domain. In this letter, a novel Multi-Instance Multi-Label learning framework based on Deep Convolutional Neural Network (MIML-DCNN) is proposed to automatically recognize the overlapping LPI radar signals,which is trained by single type of signals only. The framework handles signals in an end-to-end manner that is integrated with a well-designed instance generation module, a sophisticated MIML classifier, and an adaptive threshold calibration. Through comprehensive experiments on simulated overlapping signals with four different modulation types, we prove that the proposed framework identifies each individual signal type precisely in the presence of overlapping signals, and is also robust to variation of the signal-to-noise ratio (SNR) and power ratio conditions.

Pulse position modulation‐ultra wideband (PPM–UWB) communication signal is hard to detect and sample directly, owing to its ultra‐low power spectral density and wide bandwidth. There are already some researches on using analogue‐to‐information converter (AIC) technology and compressed sensing (CS) theory to under‐sample and detect PPM‐UWB communication signal, utilising its sparseness in time domain. However, greedy algorithm lacks of restriction on sparseness of reconstructed vector, while common restrictions on sparseness (e.g. convex optimisation) has high computational complexity. To solve these problems, a combinatorial optimisation method is proposed in this study to detect PPM–UWB communication signal based on CS and AIC. Reconstruction error and sparseness of reconstructed vector are restricted by l 2 ‐ and l p ‐norms, respectively. l p ‐norm (0 < p < 1), which is a non‐convex function, has stricter restriction on sparseness than l 1 ‐norm. Meanwhile, the steepest descent method is adopted for l p ‐norm optimisation, which can rapidly converge to objective values. Proposed method has more comprehensive restriction than greedy algorithm and convex optimisation, while maintain low complexity in computation as greedy algorithm. Numerical experiments demonstrate the validity of proposed method.

In the soft-input soft-output Viterbi algorithm (SOVA), the log-likelihood ratio (LLR) of each bit is determined by the minimum metric difference between the ML path and its competitive paths. This paper proposes to trim large metric differences in order to reduce the complexity of SOVA. By trimming the metric differences, only a small number of backtracking operations are carried out, while many LLRs may be omitted as the result of the lack of metric differences. By revealing the relationship among neighboring LLRs, the omitted LLRs are estimated from its neighoring LLRs as well as intrinsic information. The extrinsic information transfer chart analysis demonstrates that the proposed algorithm has similar convergence behavior as the Log-MAP algorithm, if the trimming factor M is moderate. Other analyses verify that our approach provides good LLR quality with only at most 1/M backtracking operations of SOVA. Simulation results show that it outperforms SOVA and performs as well as its variants and the Log-MAP algorithm.

The weighted bit-reliability-based (wBRB) algorithm for nonbinary LDPC codes suffers certain loss of symbol-reliability. Thus, this paper enhances its soft-decision version by passing multiple symbol-reliability instead of bit-reliability. Furthermore, it demonstrates that plurality robustly indicates symbol-reliability of extrinsic information-sums. Thus, this paper enhances the hard-decision version by introducing symbol-reliability from plurality. Analysis results show that these two enhanced decoding algorithms significantly outperform the wBRB algorithm with reasonable overhead.

This paper presents two packings derived from balanced incomplete block designs to construct quasi-cyclic spatially coupled LDPC convolutional codes (SC-LDPC-CCs). The construction gives time-invariant codes, since the sub-blocks corresponding to each time instant of the parity-check matrix are identical. Moreover, it provides flexible design rates and constraint lengths. Simulation results show that the proposed packing-based SC-LDPC-CCs outperform the existing time-invariant codes and perform closely to the time-varying protograph-based codes.