EvoSherlock: Towards Agentic Lifelong Evolution for Unseen Long-Tailed Security-Critical Events in Videos
本文提出EvoSherlock方法，通过因果增强和自主控制器解决长尾安全关键事件的持续分类与时间定位问题，应对数据稀缺和事件间干扰挑战。
0 citationsRead paper
Selected work
Representative Papers
Non-GRS type MDS and AMDS codes from extended TGRS codes
This work proposes a new class of extended twisted generalized Reed–Solomon (TGRS) codes to broaden the design space of optimal and near-optimal error-correcting codes. It systematically investigates the necessary and sufficient conditions under which these codes attain maximum distance separable (MDS) or almost MDS (AMDS) properties. Leveraging algebraic coding theory and equivalence analysis, the study rigorously establishes—for the first time—that the constructed codes are non-equivalent to classical Reed–Solomon (GRS) codes for specific parameter choices. Furthermore, it precisely determines their covering radii and deep holes. By providing explicit constructions of novel non-GRS MDS and AMDS codes along with clear parameter criteria, this research significantly enriches the theoretical foundations and practical resources available for high-performance error correction.
0 citationsRead paper
Robust and Well-conditioned Sparse Estimation for High-dimensional Covariance Matrices
Robust sparse estimation of high-dimensional covariance matrices faces three interrelated challenges: difficulty in guaranteeing positive definiteness, sparsity degradation due to post-hoc corrections, and uncontrolled condition numbers. Method: We propose the first method that explicitly incorporates a condition-number constraint into a robust adaptive thresholding framework. Using convex optimization and a provably convergent alternating direction algorithm, our approach jointly ensures positive definiteness, sparsity, and numerical stability. Contribution/Results: We establish theoretical minimax optimal convergence rate under the Frobenius norm. Experiments on both synthetic and real-world datasets demonstrate that our estimator consistently yields positive definite, sparse, and well-conditioned (low condition number) covariance matrices. Its numerical stability matches or surpasses that of eigenvalue truncation, while requiring fewer hyperparameters and offering greater practical utility.
0 citationsRead paper
$L_1$-norm Regularized Indefinite Kernel Logistic Regression
To address the challenge of simultaneously achieving sparsity, interpretability, and generalization in indefinite kernel logistic regression (IKLR), this paper introduces the $L_1$-norm regularization into the IKLR framework for the first time, yielding the Sparse Indefinite Kernel Logistic Regression (S-IKLR) model. To tackle the resulting nonsmooth and nonconvex optimization problem, we propose an efficient proximal linearization-based algorithm with theoretical convergence guarantees. S-IKLR leverages the expressive power of indefinite kernels to capture complex data structures while enforcing sparsity via $L_1$ regularization, thereby substantially reducing the number of nonzero parameters. Extensive experiments on multiple benchmark datasets demonstrate that S-IKLR achieves superior classification accuracy compared to state-of-the-art IKLR and sparse kernel methods. Moreover, it attains 30–60% higher model sparsity, leading to significantly improved interpretability and generalization performance.
0 citationsRead paper
Lightweight Shrimp Disease Detection Research Based on YOLOv8n
To address the low efficiency and insufficient accuracy of disease detection in shrimp aquaculture—leading to substantial economic losses—this paper proposes a lightweight YOLOv8n-based model. The method introduces three key innovations: (1) an RLDD detection head, (2) a C2f-EMCM feature fusion module, and (3) an enhanced SegNext_Attention self-attention mechanism, collectively improving multi-scale lesion feature representation while reducing computational overhead. Experiments on a custom shrimp disease dataset and the URPC2020 benchmark demonstrate that the proposed model reduces parameter count by 32.3%, achieves an mAP@0.5 of 92.7% (+3.0% improvement), and attains a +4.1% mAP@0.5 gain on URPC2020—outperforming state-of-the-art lightweight YOLO variants. The approach delivers an efficient, robust solution for intelligent disease identification in aquaculture.
0 citationsRead paper
Recent publications
Latest Papers
EvoSherlock: Towards Agentic Lifelong Evolution for Unseen Long-Tailed Security-Critical Events in Videos
本文提出EvoSherlock方法，通过因果增强和自主控制器解决长尾安全关键事件的持续分类与时间定位问题，应对数据稀缺和事件间干扰挑战。
0 citationsRead paper
Non-GRS type MDS and AMDS codes from extended TGRS codes
This work proposes a new class of extended twisted generalized Reed–Solomon (TGRS) codes to broaden the design space of optimal and near-optimal error-correcting codes. It systematically investigates the necessary and sufficient conditions under which these codes attain maximum distance separable (MDS) or almost MDS (AMDS) properties. Leveraging algebraic coding theory and equivalence analysis, the study rigorously establishes—for the first time—that the constructed codes are non-equivalent to classical Reed–Solomon (GRS) codes for specific parameter choices. Furthermore, it precisely determines their covering radii and deep holes. By providing explicit constructions of novel non-GRS MDS and AMDS codes along with clear parameter criteria, this research significantly enriches the theoretical foundations and practical resources available for high-performance error correction.
0 citationsRead paper
Robust and Well-conditioned Sparse Estimation for High-dimensional Covariance Matrices
Robust sparse estimation of high-dimensional covariance matrices faces three interrelated challenges: difficulty in guaranteeing positive definiteness, sparsity degradation due to post-hoc corrections, and uncontrolled condition numbers. Method: We propose the first method that explicitly incorporates a condition-number constraint into a robust adaptive thresholding framework. Using convex optimization and a provably convergent alternating direction algorithm, our approach jointly ensures positive definiteness, sparsity, and numerical stability. Contribution/Results: We establish theoretical minimax optimal convergence rate under the Frobenius norm. Experiments on both synthetic and real-world datasets demonstrate that our estimator consistently yields positive definite, sparse, and well-conditioned (low condition number) covariance matrices. Its numerical stability matches or surpasses that of eigenvalue truncation, while requiring fewer hyperparameters and offering greater practical utility.
0 citationsRead paper
$L_1$-norm Regularized Indefinite Kernel Logistic Regression
To address the challenge of simultaneously achieving sparsity, interpretability, and generalization in indefinite kernel logistic regression (IKLR), this paper introduces the $L_1$-norm regularization into the IKLR framework for the first time, yielding the Sparse Indefinite Kernel Logistic Regression (S-IKLR) model. To tackle the resulting nonsmooth and nonconvex optimization problem, we propose an efficient proximal linearization-based algorithm with theoretical convergence guarantees. S-IKLR leverages the expressive power of indefinite kernels to capture complex data structures while enforcing sparsity via $L_1$ regularization, thereby substantially reducing the number of nonzero parameters. Extensive experiments on multiple benchmark datasets demonstrate that S-IKLR achieves superior classification accuracy compared to state-of-the-art IKLR and sparse kernel methods. Moreover, it attains 30–60% higher model sparsity, leading to significantly improved interpretability and generalization performance.
0 citationsRead paper
Lightweight Shrimp Disease Detection Research Based on YOLOv8n
To address the low efficiency and insufficient accuracy of disease detection in shrimp aquaculture—leading to substantial economic losses—this paper proposes a lightweight YOLOv8n-based model. The method introduces three key innovations: (1) an RLDD detection head, (2) a C2f-EMCM feature fusion module, and (3) an enhanced SegNext_Attention self-attention mechanism, collectively improving multi-scale lesion feature representation while reducing computational overhead. Experiments on a custom shrimp disease dataset and the URPC2020 benchmark demonstrate that the proposed model reduces parameter count by 32.3%, achieves an mAP@0.5 of 92.7% (+3.0% improvement), and attains a +4.1% mAP@0.5 gain on URPC2020—outperforming state-of-the-art lightweight YOLO variants. The approach delivers an efficient, robust solution for intelligent disease identification in aquaculture.
0 citationsRead paper