As Audio Large Language Models (ALLMs) emerge as powerful tools for speech processing, their safety implications demand urgent attention. While considerable research has explored textual and vision safety, audio’s distinct characteristics present significant challenges. This paper first investigates: Is ALLM vulnerable to backdoor attacks exploiting acoustic triggers? In response to this issue, we introduce Hidden in the Noise (HIN), a novel backdoor attack framework designed to exploit subtle, audio-specific features. HIN applies acoustic modifications to raw audio waveforms, such as alterations to temporal dynamics and strategic injection of spectrally tailored noise. These changes introduce consistent patterns that an ALLM’s acoustic feature encoder captures, embedding robust triggers within the audio stream. To evaluate ALLM robustness against audio-feature-based triggers, we develop the AudioSafe benchmark, assessing nine distinct risk types. Extensive experiments on AudioSafe and three established safety datasets reveal critical vulnerabilities in existing ALLMs: (I) audio features like environment noise and speech rate variations achieve over 90% average attack success rate, (II) ALLMs exhibit significant sensitivity differences across acoustic features, particularly showing minimal response to volume as a trigger, and (III) poisoned sample inclusion causes only marginal loss curve fluctuations, highlighting the attack’s stealth.

With the popularity of electric vehicles (EVs), wireless power transmission (WPT) technology has become a hot research topic for next-generation battery charging technology. However, the vulnerability of wireless networks to malicious interference attacks is inherited by WPT. To alleviate the privacy and security issues of WPT, we propose a novel FedDQ, a federated deep reinforcement learning with Q-ensemble, to cope with interference attacks in EV wireless charging network environments. Federated learning protects the security privacy of EVs by training a global model that exploits the property that data and models will not be transmitted. In order to trade-off the training cost and efficiency, we introduce offline-to-online training models by pre-training the offline Q-network with pre-collected data, and the trained model serves as an initialization of the online model. Then, the online Q-network is obtained by weakening or removing the original pessimistic constraints to enhance the training speed. Secondly, we introduce the intelligent reflective surface (IRS) to enhance the security performance of WPT by modifying the IRS phase shift and amplitude to cancel the malicious interference signal. Experimental results show that our proposed FedDQ algorithm has superior performance and outperforms existing baseline methods in terms of anti-jamming metrics.

Weakly-Supervised Video Anomaly Detection (WS-VAD) primarily struggles to robustly model normal temporal dynamics and sensitively detect sparse anomalies using only video-level labels. Current methods often inadequately represent normal patterns or are too computationally costly due to complex architectures, hindering practical use. We introduce DSIB-Net, a lightweight, dual-stream Information Bottleneck (IB) based framework. Its core Dual-Output Information Bottleneck Encoder (DOIBE), driven by a Temporal Information Bottleneck Loss (TIBL), operationalizes IB. DOIBE produces a predictive representation (zpred) for dynamic modeling and a compression control vector (zcomp) for redundancy compression, with TIBL assigning them distinct functional roles. Since IB compression in DOIBE discards fine-grained appearance details, a parallel Instantaneous Mode Monitoring Path (IMP) provides crucial compensation by extracting these neglected frame-level features. An Adaptive Feature Arbitration (AFA) module then intelligently fuses these streams, dynamically weighting zpred based on IMP’s appearance cues to balance robustness and sensitivity. Experiments show DSIB-Net delivers highly competitive detection performance with minimal computational overhead, offering an efficient and practical WS-VAD solution.The code and dataset for the model are open-source and available at https://github.com/modadundun/DSIB-Net.

Named Entity Recognition (NER) aims to identify the pre-defined entities from the unstructured text. Compared with English NER, Chinese NER faces more challenges: the ambiguity problem in entity boundary recognition due to unavailable explicit delimiters between Chinese characters, and the out-of-vocabulary (OOV) problem caused by rare Chinese characters. However, two important features specific to the Chinese language are ignored by previous studies: glyphs and phonetics, which contain rich semantic information of Chinese. To overcome these issues by exploiting the linguistic potential of Chinese as a logographic language, we present MPM-CNER (short for Multi-modal Pretraining Model for Chinese NER), a model for learning multi-modal representations of Chinese semantics, glyphs, and phonetics, via four pretraining tasks: Radical Consistency Identification (RCI), Glyph Image Classification (GIC), Phonetic Consistency Identification (PCI), and Phonetic Classification Modeling (PCM). Meanwhile, a novel cross-modality attention mechanism is proposed to fuse these multimodal features for further improvement. The experimental results show that our method outperforms the state-of-the-art baseline methods on four benchmark datasets, and the ablation study also verifies the effectiveness of the pre-trained multi-modal representations.

Aiming at the “dimension disaster” problem encountered in the outlier detection of high-dimensional data, this paper uses the projection pursuit algorithm to perform non-linear dimensionality reduction on high-dimensional data by calculating the phase relationship between dimensions. According to the sample points obtained by dimensionality reduction, the LOF (Local Outlier Factor) algorithm is applied to calculate the outlier factor to obtain the relevant outlier data. In order to improve the calculation accuracy and efficiency of the LOF algorithm, clustering method is used to cut the outlier calculation data to reduce the amount of calculation. Experiments on real-world and artificial datasets, compared with the existing algorithms, demonstrated the effectiveness and efficiency of the proposed algorithm.