Few-shot multimodal industrial anomaly detection is a critical yet underexplored task, offering the ability to quickly adapt to complex industrial scenarios. In few-shot settings, insufficient training samples often fail to cover the diverse patterns present in test samples. This challenge can be mitigated by extracting structural commonality from a small number of training samples. In this paper, we propose a novel few-shot unsupervised multimodal industrial anomaly detection method based on structural commonality, CIF (Commonality In Few). To extract intra-class structural information, we employ hypergraphs, which are capable of modeling higher-order correlations, to capture the structural commonality within training samples, and use a memory bank to store this intra-class structural prior. Firstly, we design a semantic-aware hypergraph construction module tailored for single-semantic industrial images, from which we extract common structures to guide the construction of the memory bank. Secondly, we use a training-free hypergraph message passing module to update the visual features of test samples, reducing the distribution gap between test features and features in the memory bank. We further propose a hyperedge-guided memory search module, which utilizes structural information to assist the memory search process and reduce the false positive rate. Experimental results on the MVTec 3D-AD dataset and the Eyecandies dataset show that our method outperforms the state-of-the-art (SOTA) methods in few-shot settings.

In modern healthcare, emotional well-being is critical to patient recovery and overall outcomes. However, limited availability of trained professionals and time constraints often hinder the delivery of consistent emotional support. To address this gap, we propose the Embodied Emotional Care System (EECS), a comprehensive humanoid robotic framework designed to deliver personalized emotional care through an integrated, multi-layered architecture. EECS analyzes dynamic facial expressions and real-time vocal inputs to extract the patient’s emotional state and semantic information, constructs context-aware prompts processed by an LLM for reasoning, and ultimately generates empathetic dialogues synchronized with human-like facial expressions and natural body movements to address diverse emotional support needs. Experimental results show that deploying EECS on a humanoid robot significantly boosts patient engagement through real-time multimodal interaction, delivering deeper emotional support and a more human-like therapeutic experience. Furthermore, it bridges gaps in professional emotional support resources, offering a feasible pathway to improve overall healthcare quality.

Anomaly detection is critical in industrial manufacturing for ensuring product quality and improving efficiency in automated processes. The scarcity of anomalous samples limits traditional detection methods, making anomaly generation essential for expanding the data repository. However, recent generative models often produce unrealistic anomalies increasing false positives, or require real-world anomaly samples for training. In this work, we treat anomaly generation as a compositional problem and propose ComGEN, a component-aware and unsupervised framework that addresses the gap in logical anomaly generation. Our method comprises a multi-component learning strategy to disentangle visual components, followed by subsequent generation editing procedures. Disentangled text-to-component pairs, revealing intrinsic logical constraints, conduct attention-guided residual mapping and model training with iteratively matched references across multiple scales. Experiments on the MVTecLOCO dataset confirm the efficacy of ComGEN, achieving the best AUROC score of$\mathbf{9 1. 2 \%}$. Additional experiments on the real-world scenario of Diesel Engine and widelyused MVTecAD dataset demonstrate significant performance improvements when integrating simulated anomalies generated by ComGEN into automated production workflows.

Anomaly detection and localization in automated industrial manufacturing can significantly enhance production efficiency and product quality. Existing methods are capable of detecting surface defects in pre-defined or controlled imaging environments. However, accurately detecting workpiece defects in complex and unstructured industrial environments with varying views, poses and illumination remains challenging. We propose a novel anomaly detection and localization method specifically designed to handle inputs with perturbative patterns. Our approach introduces a new framework based on a collaborative distillation heterogeneous teacher network (HetNet), an adaptive local-global feature fusion module, and a local multivariate Gaussian noise generation module. HetNet can learn to model the complex feature distribution of normal patterns using limited information about local disruptive changes. We conducted extensive experiments on mainstream benchmarks. HetNet demonstrates superior performance with approximately 10% improvement across all evaluation metrics on MSC-AD under industrial conditions, while achieving state-of-the-art results on other datasets, validating its resilience to environmental fluctuations and its capability to enhance the reliability of industrial anomaly detection systems across diverse scenarios. Tests in real-world environments further confirm that HetNet can be effectively integrated into production lines to achieve robust and real-time anomaly detection. Codes, images and videos are published on the project website at: https://zihuatanejoyu.github.io/HetNet/

This paper proposes a novel radiation-hardened high-reliability SRAM cell, namely SHRCO, with 12 transistors for robust value storage as well as 6 transistors for parallel access operations. Using separated and error-interceptive feedback paths, the proposed cell has a complete self-recoverability from single-node upset (SNUs) at all single nodes and an excellent self-recoverability from double-node upsets (DNUs) at a part of node pairs. In addition, the proposed cell has superior access operation speed due to the inclusion of extra parallel access transistors. Simulation results show that the proposed cell has the largest number of node pairs that can self-recover from DNUs. Moreover, compared to the existing radiation-hardened SRAM cells, the proposed cell saves 28% of read time and 3% of write time on average.

Anomaly detection aims to separate anomalies from normal samples, and the pretrained network is promising for anomaly detection. However, adapting the pretrained features would be confronted with the risk of pattern collapse when finetuning on one-class training data. In this paper, we propose an anomaly detection framework called constrained adaptive projection with pretrained features (CAP). Combined with pretrained features, a simple linear projection head applied on a specific input and its k most similar pretrained normal representations is designed for feature adaptation, and a reformed self-attention is leveraged to mine the inner-relationship among one-class semantic features. A loss function is proposed to avoid potential pattern collapse. Concretely, it considers the similarity between a specific data and its corresponding adaptive normal representation, and incorporates a constraint term slightly aligning pretrained and adaptive spaces. Our method achieves state-of-the-art anomaly detection performance on semantic anomaly detection and sensory anomaly detection benchmarks including 96.5% AUROC on CIFAR-100 dataset, 97.0% AUROC on CIFAR-10 dataset and 89.9% AUROC on MvTec dataset.

In this paper we propose a new approach of two-dimensional frame-and-feature weighted Viterbi decoding performed at the recognizer back-end for robust speech recognition. A new SVM-based frame weighting approach is proposed considering the energy distribution and harmonicity of the frame. The feature weighting is based on a previously proposed approach using an entropy measure considering confusion between phoneme classes. These two different weighting schemes on the two different dimensions are then properly integrated in Viterbi decoding in this paper. Extensive experiments performed with the Aurora 4 testing environment showed significant improvements.

Most stream data classification algorithms apply the supervised learning strategy which requires massive labeled data. Such approaches are impractical since labeled data are usually hard to obtain in reality. In this paper, we build a clustering feature decision tree model, CFDT, from data streams having both unlabeled and a small number of labeled examples. CFDT applies a micro-clustering algorithm that scans the data only once to provide the statistical summaries of the data for incremental decision tree induction. Micro-clusters also serve as classifiers in tree leaves to improve classification accuracy and reinforce the any-time property. Our experiments on synthetic and real-world datasets show that CFDT is highly scalable for data streams while generating high classification accuracy with high speed.

Hard real-time systems always choose not to use garbage collection in order to avoid its unpredictable executions. Much effort has been expended trying to build predictable garbage collectors which can provide both temporal and spatial guarantees. Unfortunately, most existing work leads to systems that cannot easily achieve a balance between temporal and spatial performances. Moreover, the scheduling of garbage collectors has not been integrated into modern real-time scheduling frameworks, which makes the benefits provided by the advancement of scheduling techniques very difficult to obtain. This paper argues that the existing design criteria for real-time garbage collectors do not reflect the unique requirements of flexible hard real-time systems. As a part of our design criteria, a new performance indicator is proposed to describe the capability of a real-time garbage collector to achieve a better balance between temporal and spatial performances. A hybrid garbage collection algorithm is designed accordingly which also uses dual priority scheduling algorithm to reclaim spare capacity while guaranteeing deadlines.

Real-time garbage collection algorithms are usually criticised for their high memory requirements. Even when consuming nearly 50% of CPU time, some garbage collectors ask for at least twice the memory as really needed. This paper explores the fundamental reason for this problem and proposes a new performance indicator for the evaluation of real-time garbage collection algorithms. Use of this performance indicator motivates an algorithm that combines both reference counting and mark-and-sweep techniques. In the presence of our collector, a garbage collected hard real-time system can achieve the correct balance of time-space tradeoff with less effort. In order to provide both temporal and spatial guarantees needed by a hard real-time application, an offline analysis is developed and integrated into the response time analysis framework. Moreover, the use of dual priority scheduling of the garbage collection tasks allows spare capacity in the system to be reclaimed whilst guaranteeing deadlines

In this paper, we propose an approach to integrate a hybrid garbage collection algorithm (a combination of reference counting and mark-and-sweep techniques) into the current response time analysis framework for real-time systems. Instead of collecting garbage incrementally, we put most GC work into a periodic real-time thread, namely the GC thread, which is scheduled according to the dual priority scheduling method. More importantly, we can perform schedulability analysis (response time analysis) for all the real-time threads including the GC thread.