We pioneer a learning-based single-point prompt paradigm for infrared small target label generation (IRSTLG) to lobber annotation burdens. Unlike previous clustering-based methods, our intuition is that point-guided mask generation just requires one more prompt than target detection, i.e., IRSTLG can be treated as an infrared small target detection (IRSTD) with the location hint. Therefore, we propose an elegant yet effective Energy-Double-Guided Single-point Prompt (EDGSP) framework, aiming to adeptly transform a coarse IRSTD network into a refined label generation method. Specifically, EDGSP comprises three key modules: 1) target energy initialization (TEI), which establishes a foundational outline to streamline the mapping process for effective shape evolution, 2) double prompt embedding (DPE) for rapidly localizing interesting regions and reinforcing high-resolution individual edges to avoid label adhesion, and 3) bounding box-based matching (BBM) for eliminating false masks via considering comprehensive cluster boundary conditions to obtain a reliable output. In this way, pseudo labels generated by three backbones equipped with our EDGSP achieve 100% object-level probability of detection (Pd) and 0% false-alarm rate (Fa) on SIRST, NUDT-SIRST, and IRSTD-1k datasets, with a pixel-level intersection over union (IoU) improvement of 13.28% over state-of-the-art (SOTA) label generation methods. Further applying our inferred masks to train detection models, EDGSP, for the first time, enables a single-point-generated pseudo mask to surpass the manual labels. Even with coarse single-point annotations, it still achieves 99.5% performance of full labeling. Code is available at this https URL

The technique of transforming voices in order to hide the real identity of a speaker is called voice disguise, among which automatic voice disguise (AVD) by modifying the spectral and temporal characteristics of voices with miscellaneous algorithms are easily conducted with softwares accessible to the public. AVD has posed great threat to both human listening and automatic speaker verification (ASV). In this paper, we have found that ASV is not only a victim of AVD but could be a tool to beat some simple types of AVD. Firstly, three types of AVD, pitch scaling, vocal tract length normalization (VTLN) and voice conversion (VC), are introduced as representative methods. State-of-the-art ASV methods are subsequently utilized to objectively evaluate the impact of AVD on ASV by equal error rates (EER). Moreover, an approach to restore disguised voice to its original version is proposed by minimizing a function of ASV scores w.r.t. restoration parameters. Experiments are then conducted on disguised voices from Voxceleb, a dataset recorded in real-world noisy scenario. The results have shown that, for the voice disguise by pitch scaling, the proposed approach obtains an EER around 7% comparing to the 30% EER of a recently proposed baseline using the ratio of fundamental frequencies. The proposed approach generalizes well to restore the disguise with nonlinear frequency warping in VTLN by reducing its EER from 34.3% to 18.5%. However, it is difficult to restore the source speakers in VC by our approach, where more complex forms of restoration functions or other paralinguistic cues might be necessary to restore the nonlinear transform in VC. Finally, contrastive visualization on ASV features with and without restoration illustrate the role of the proposed approach in an intuitive way.

Deep neural networks are vulnerable to adversarial examples, which can fool deep models by adding subtle perturbations. Although existing attacks have achieved promising results, it still leaves a long way to go for generating transferable adversarial examples under the black-box setting. To this end, this paper proposes to improve the transferability of adversarial examples, and applies dual-stage feature-level perturbations to an existing model to implicitly create a set of diverse models. Then these models are fused by the longitudinal ensemble during the iterations. The proposed method is termed Dual-Stage Network Erosion (DSNE). We conduct comprehensive experiments both on non-residual and residual networks, and obtain more transferable adversarial examples with the computational cost similar to the state-of-the-art method. In particular, for the residual networks, the transferability of the adversarial examples can be significantly improved by biasing the residual block information to the skip connections. Our work provides new insights into the architectural vulnerability of neural networks and presents new challenges to the robustness of neural networks.

Edge intelligence is anticipated to underlay the pathway to connected intelligence for 6G networks, but the organic confluence of edge computing and artificial intelligence still needs to be carefully treated. To this end, this article discusses the concepts of edge intelligence from the semantic cognitive perspective. Two instructive theoretical models for edge semantic cognitive intelligence (ESCI) are first established. Afterwards, the ESCI framework orchestrating deep learning with semantic communication is discussed. Two representative applications are present to shed light on the prospect of ESCI in 6G networks. Some open problems are finally listed to elicit the future research directions of ESCI.

Human skeleton data has received increasing attention in action recognition due to its background robustness and high efficiency. In skeleton-based action recognition, graph convolutional network (GCN) has become the mainstream method. This paper analyzes the fundamental factor for GCN-based models -- the adjacency matrix. We notice that most GCN-based methods conduct their adjacency matrix based on the human natural skeleton structure. Based on our former work and analysis, we propose that the human natural skeleton structure adjacency matrix is not proper for skeleton-based action recognition. We propose a new adjacency matrix that abandons all rigid neighbor connections but lets the model adaptively learn the relationships of joints. We conduct extensive experiments and analysis with a validation model on two skeleton-based action recognition datasets (NTURGBD60 and FineGYM). Comprehensive experimental results and analysis reveals that 1) the most widely used human natural skeleton structure adjacency matrix is unsuitable in skeleton-based action recognition; 2) The proposed adjacency matrix is superior in model performance, noise robustness and transferability.

In recent years, the field of intelligent transportation systems (ITS) has achieved remarkable success, which is mainly due to the large amount of available annotation data. However, obtaining these annotated data has to afford expensive costs in reality. Therefore, a more realistic strategy is to leverage semi-supervised learning (SSL) with a small amount of labeled data and a large amount of unlabeled data. Typically, semantic consistency regularization and the two-stage learning methods of decoupling feature extraction and classification have been proven effective. Nevertheless, representation learning only limited to semantic consistency regularization may not guarantee the separation or discriminability of representations of samples with different semantics; due to the inherent limitations of the two-stage learning methods, the extracted features may not match the specific downstream tasks. In order to deal with the above drawbacks, this paper proposes an end-to-end deep semi-supervised learning double contrast of semantic and feature, which extracts effective tasks specific discriminative features by contrasting the semantics/features of positive and negative augmented samples pairs. Moreover, we leverage information theory to explain the rationality of double contrast of semantics and features and slack mutual information to contrastive loss in a simpler way. Finally, the effectiveness of our method is verified in benchmark datasets.

The "search for extraterrestrial intelligence" (SETI) commensal surveys aim to scan the sky to find possible technosignatures from the extraterrestrial intelligence (ETI). The mitigation of radio frequency interference (RFI) is an important step, especially for the most sensitive Five-hundred-meter Aperture Spherical radio Telescope (FAST), which can detect more weak RFI. In this paper, we propose several new techniques for RFI mitigation, and use our procedure to search for ETI signals from the archival data of FAST's first SETI commensal survey. We detect the persistent narrowband RFI by setting a threshold of the signals' sky separation, and detect the drifting RFI (and potentially other types of RFI) using the Hough transform. We also use the clustering algorithms to remove more RFI and select candidates. The results of our procedure are compared to the earlier work on the same FAST data. We find that our methods, though relatively simpler in computation, remove more RFI (99.9912% compared to 99.9063% in the earlier work), but preserve the simulated ETI signals except those (5.1%) severely affected by the RFI. We also report more interesting candidate signals, about a dozen of which are new candidates that are not previously reported. In addition, we find that the proposed Hough transform method, with suitable parameters, also has the potential to remove the broadband RFI. We conclude that our methods can effectively remove the vast majority of the RFI while preserving and finding the candidate signals that we are interested in.

The radio environment map (REM) visually displays the spectrum information over the geographical map and plays a significant role in monitoring, management, and security of spectrum resources.In this paper, we present an efficient 3D REM construction scheme based on the sparse Bayesian learning (SBL), which aims to recovery the accurate REM with limited and optimized sampling data.In order to reduce the number of sampling sensors, an efficient sparse sampling method for unknown scenarios is proposed. For the given construction accuracy and the priority of each location, the quantity and sampling locations can be jointly optimized.With the sparse sampled data, by mining the spectrum situation sparsity and channel propagation characteristics, an SBL-based spectrum data hierarchical recovery algorithm is developed to estimate the missing data of unsampled locations.Finally, the simulated 3D REM data in the campus scenario are used to verify the proposed methods as well as to compare with the state-of-the-art. We also analyze the recovery performance and the impact of different parameters on the constructed REMs. Numerical results demonstrate that the proposed scheme can ensure the construction accuracy and improve the computational efficiency under the low sampling rate.

In this paper, we propose a deep learning (DL)-based task-driven spectrum prediction framework, named DeepSPred. The DeepSPred comprises a feature encoder and a task predictor, where the encoder extracts spectrum usage pattern features, and the predictor configures different networks according to the task requirements to predict future spectrum. Based on the Deep- SPred, we first propose a novel 3D spectrum prediction method combining a flow processing strategy with 3D vision Transformer (ViT, i.e., Swin) and a pyramid to serve possible applications such as spectrum monitoring task, named 3D-SwinSTB. 3D-SwinSTB unique 3D Patch Merging ViT-to-3D ViT Patch Expanding and pyramid designs help the model accurately learn the potential correlation of the evolution of the spectrogram over time. Then, we propose a novel spectrum occupancy rate (SOR) method by redesigning a predictor consisting exclusively of 3D convolutional and linear layers to serve possible applications such as dynamic spectrum access (DSA) task, named 3D-SwinLinear. Unlike the 3D-SwinSTB output spectrogram, 3D-SwinLinear projects the spectrogram directly as the SOR. Finally, we employ transfer learning (TL) to ensure the applicability of our two methods to diverse spectrum services. The results show that our 3D-SwinSTB outperforms recent benchmarks by more than 5%, while our 3D-SwinLinear achieves a 90% accuracy, with a performance improvement exceeding 10%.

Competitive Co-evolutionary Algorithms (CCEAs) are often hampered by complex dynamics like intransitivity and the Red Queen effect, leading to unstable convergence. To counter these challenges, this paper introduces the Marker Gene Method (MGM), a framework that establishes stability by using a 'marker gene' as a dynamic benchmark and an adaptive weighting mechanism to balance exploration and exploitation. We provide rigorous mathematical proofs demonstrating that MGM creates strong attractors near Nash Equilibria within the Strictly Competitive Game framework. Empirically, MGM demonstrates its efficacy across a spectrum of challenges: it stabilizes the canonical Rock-Paper-Scissors game, significantly improves the performance of C-RMOEA/D on ZDT benchmarks, and, when augmented with a Memory Pool (MP) extension, it successfully tames the notoriously pathological Shapley Biased Game. This work presents a theoretically sound and empirically validated framework that substantially enhances the stability and robustness of CCEAs in complex competitive environments.

We investigate quantum phase transitions and quantum coherence in a quantum compass chain under an alternating transverse magnetic field. The model can be analytically solved by the Jordan-Wigner transformation and this solution shows that it is equivalent to a two-component one-dimensional (1D) Fermi gas on a lattice. We explore mutual effects of the staggered magnetic interaction and multi-site interactions on the energy spectra and analyze the ground state phase diagram. We use quantum coherence measures to identify the quantum phase transitions. Our results show that $l_1$ norm of coherence fails to detect faithfully the quantum critical points separating a gapped phase from a gapless phase, which can be pinpointed exactly by relative entropy of coherence. Jensen-Shannon divergence is somewhat obscure at exception points. We also propose an experimental realization of such a 1D system using superconducting quantum circuits.

With the increasingly complex and changeable electromagnetic environment, wireless communication systems are facing jamming and abnormal signal injection, which significantly affects the normal operation of a communication system. In particular, the abnormal signals may emulate the normal signals, which makes it very challenging for abnormal signal recognition. In this paper, we propose a new abnormal signal recognition scheme, which combines time-frequency analysis with deep learning to effectively identify synthetic abnormal communication signals. Firstly, we emulate synthetic abnormal communication signals including seven jamming patterns. Then, we model an abnormal communication signals recognition system based on the communication protocol between the transmitter and the receiver. To improve the performance, we convert the original signal into the time-frequency spectrogram to develop an image classification algorithm. Simulation results demonstrate that the proposed method can effectively recognize the abnormal signals under various parameter configurations, even under low signal-to-noise ratio (SNR) and low jamming-to-signal ratio (JSR) conditions.

A cavity quantum electrodynamical (QED) system beyond the strong-coupling regime is expected to exhibit intriguing quantum phenomena. Here we report a direct measurement of the photon-dressed qubit transition frequencies up to four photons by harnessing the same type of state transitions in an ultrastrongly coupled circuit-QED system realized by inductively coupling a superconducting flux qubit to a coplanar-waveguide resonator. This demonstrates a convincing observation of the photon-dressed Bloch-Siegert shift in the ultrastrongly coupled quantum system. Moreover, our results show that the photon-dressed Bloch-Siegert shift becomes more pronounced as the photon number increases, which is a characteristic of the quantum Rabi model.