Selected work

Representative Papers

Dual Spatial-Temporal Attribution: Architecture-Aligned Post-Hoc Explainability for Recurrent Graph Anomaly Detection

Aug 12, 2026

This work addresses the limited interpretability of dynamic graph anomaly detection models, which hinders their applicability in audit and trustworthy decision-making scenarios. The authors propose X-AddGraph, the first post-hoc explainability method tailored for the AddGraph framework. By introducing an architecture-aligned, three-component attribution mechanism, X-AddGraph jointly explains spatial structure, short-term temporal attention, and long-term recurrent states without compromising the original model’s performance. The approach integrates gradient-based attribution, attention weight inspection, and hidden state backtracking to deliver precise, reproducible explanations. Evaluated on the UCI Message dataset, the method achieves a reproduced AUC of 0.8705, with long-term temporal attribution significantly outperforming random baselines (0.127 vs. 0.074), thereby offering both high fidelity and comprehensive interpretability.

0 citationsRead paper

Hybridizing a Grouping Metaheuristic with Reinforcement Learning for the One-Dimensional Bin Packing Problem

Jul 02, 2026

This work addresses the one-dimensional bin packing problem (1D-BPP), an NP-hard combinatorial optimization challenge, by proposing a novel approach that integrates reinforcement learning with a hybrid grouping genetic algorithm (HGGA). A Q-learning-based controller is designed to dynamically schedule eight high-level operators—including BPCX crossover, light and heavy mutation, Martello-Toth local search, and population restart—based on an eight-dimensional state encoding, thereby achieving the first adaptive orchestration of multiple optimization operators. Experimental results demonstrate that the proposed method attains an average optimality gap of merely 0.95% on standard benchmark instances while reducing computation time from 64.22 seconds to 1.29 seconds—nearly a 50-fold speedup—significantly outperforming classical heuristics such as First-Fit Decreasing (FFD) while preserving the solution quality of HGGA.

0 citationsRead paper

GPU Parallelization Strategies for Forward and Backward Propagation in Shallow Neural Networks: A CUDA-Based Comparative Study

Jun 29, 2026

This work addresses the inefficiency of forward and backward propagation in shallow neural networks on GPUs, which stems from memory access bottlenecks. To mitigate this issue, the authors propose a three-tier CUDA-based optimization strategy that integrates tiled shared-memory padding, pre-transposition of weight matrices, and fusion of matrix multiplication (MatMul) with ReLU activation operators. By enhancing memory access patterns, minimizing global memory round trips, and increasing computational density, the approach achieves a 1.41× speedup on an NVIDIA Tesla T4 GPU—reducing execution time from 21.0 seconds to 14.8 seconds. The optimized implementation significantly outperforms both serial CPU and OpenMP-parallelized baselines, demonstrating the effectiveness of the proposed techniques in accelerating training for shallow neural networks.

0 citationsRead paper

Hybrid Metaheuristic Combining the Dragonfly Algorithm and Tabu Search for the Traveling Salesman Problem

Jun 08, 2026

This study addresses the Traveling Salesman Problem (TSP), a classic NP-hard combinatorial optimization challenge, by proposing a high-order relay hybrid solving framework. The approach innovatively integrates the swarm intelligence-based Dragonfly Algorithm (DA) with the memory-driven Tabu Search (TS): DA first performs global exploration to generate high-quality initial solutions, which are subsequently refined through local search via TS. Coupled with a systematic grid-search parameter tuning strategy, the proposed framework demonstrates superior performance on standard TSPLIB instances compared to standalone DA or TS, and outperforms classical metaheuristics such as Genetic Algorithms and Ant Colony Optimization. The results indicate significant improvements in both solution quality and robustness.

0 citationsRead paper

AgrI Challenge: A Data-Centric AI Competition for Cross-Team Validation in Agricultural Vision

Mar 07, 2026

This study addresses the poor generalization of agricultural vision models in real-world field conditions, often caused by distribution shifts between training and deployment data—a challenge overlooked by existing benchmarks that neglect the impact of data collection diversity. To this end, the authors construct a heterogeneous dataset comprising 50,673 tree images independently collected by 12 teams and propose a cross-team validation (CTV) evaluation framework, employing TOTO (Train-One, Test-One) and LOTO (Leave-One-Team-Out) protocols to systematically assess cross-source generalization. Experiments reveal that models trained on a single source suffer accuracy drops of up to 16.20%, whereas multi-source collaborative training reduces the generalization gap to within 1.78%. This work establishes the first data-centric agricultural AI competition framework, highlighting the critical role of data collection diversity in model robustness, and releases the first public agricultural vision benchmark dataset.

0 citationsRead paper

Recent publications

Latest Papers

Dual Spatial-Temporal Attribution: Architecture-Aligned Post-Hoc Explainability for Recurrent Graph Anomaly Detection

Aug 12, 2026

This work addresses the limited interpretability of dynamic graph anomaly detection models, which hinders their applicability in audit and trustworthy decision-making scenarios. The authors propose X-AddGraph, the first post-hoc explainability method tailored for the AddGraph framework. By introducing an architecture-aligned, three-component attribution mechanism, X-AddGraph jointly explains spatial structure, short-term temporal attention, and long-term recurrent states without compromising the original model’s performance. The approach integrates gradient-based attribution, attention weight inspection, and hidden state backtracking to deliver precise, reproducible explanations. Evaluated on the UCI Message dataset, the method achieves a reproduced AUC of 0.8705, with long-term temporal attribution significantly outperforming random baselines (0.127 vs. 0.074), thereby offering both high fidelity and comprehensive interpretability.

0 citationsRead paper

Hybridizing a Grouping Metaheuristic with Reinforcement Learning for the One-Dimensional Bin Packing Problem

Jul 02, 2026

This work addresses the one-dimensional bin packing problem (1D-BPP), an NP-hard combinatorial optimization challenge, by proposing a novel approach that integrates reinforcement learning with a hybrid grouping genetic algorithm (HGGA). A Q-learning-based controller is designed to dynamically schedule eight high-level operators—including BPCX crossover, light and heavy mutation, Martello-Toth local search, and population restart—based on an eight-dimensional state encoding, thereby achieving the first adaptive orchestration of multiple optimization operators. Experimental results demonstrate that the proposed method attains an average optimality gap of merely 0.95% on standard benchmark instances while reducing computation time from 64.22 seconds to 1.29 seconds—nearly a 50-fold speedup—significantly outperforming classical heuristics such as First-Fit Decreasing (FFD) while preserving the solution quality of HGGA.

0 citationsRead paper

GPU Parallelization Strategies for Forward and Backward Propagation in Shallow Neural Networks: A CUDA-Based Comparative Study

Jun 29, 2026

This work addresses the inefficiency of forward and backward propagation in shallow neural networks on GPUs, which stems from memory access bottlenecks. To mitigate this issue, the authors propose a three-tier CUDA-based optimization strategy that integrates tiled shared-memory padding, pre-transposition of weight matrices, and fusion of matrix multiplication (MatMul) with ReLU activation operators. By enhancing memory access patterns, minimizing global memory round trips, and increasing computational density, the approach achieves a 1.41× speedup on an NVIDIA Tesla T4 GPU—reducing execution time from 21.0 seconds to 14.8 seconds. The optimized implementation significantly outperforms both serial CPU and OpenMP-parallelized baselines, demonstrating the effectiveness of the proposed techniques in accelerating training for shallow neural networks.

0 citationsRead paper

Hybrid Metaheuristic Combining the Dragonfly Algorithm and Tabu Search for the Traveling Salesman Problem

Jun 08, 2026

This study addresses the Traveling Salesman Problem (TSP), a classic NP-hard combinatorial optimization challenge, by proposing a high-order relay hybrid solving framework. The approach innovatively integrates the swarm intelligence-based Dragonfly Algorithm (DA) with the memory-driven Tabu Search (TS): DA first performs global exploration to generate high-quality initial solutions, which are subsequently refined through local search via TS. Coupled with a systematic grid-search parameter tuning strategy, the proposed framework demonstrates superior performance on standard TSPLIB instances compared to standalone DA or TS, and outperforms classical metaheuristics such as Genetic Algorithms and Ant Colony Optimization. The results indicate significant improvements in both solution quality and robustness.

0 citationsRead paper

AgrI Challenge: A Data-Centric AI Competition for Cross-Team Validation in Agricultural Vision

Mar 07, 2026

This study addresses the poor generalization of agricultural vision models in real-world field conditions, often caused by distribution shifts between training and deployment data—a challenge overlooked by existing benchmarks that neglect the impact of data collection diversity. To this end, the authors construct a heterogeneous dataset comprising 50,673 tree images independently collected by 12 teams and propose a cross-team validation (CTV) evaluation framework, employing TOTO (Train-One, Test-One) and LOTO (Leave-One-Team-Out) protocols to systematically assess cross-source generalization. Experiments reveal that models trained on a single source suffer accuracy drops of up to 16.20%, whereas multi-source collaborative training reduces the generalization gap to within 1.78%. This work establishes the first data-centric agricultural AI competition framework, highlighting the critical role of data collection diversity in model robustness, and releases the first public agricultural vision benchmark dataset.

0 citationsRead paper