SeaCausal-FL: Federated Fuzzy Causal Learning for Maritime IoT Fault Diagnosis and Counterfactual Reasoning
为解决海洋物联网中因数据分布和条件变化导致的故障诊断难题，提出SeaCausal-FL框架，结合联邦学习与模糊因果推理方法，提高了诊断准确性。
0 citationsRead paper
Selected work
Representative Papers
FuDU: A Fuzzy Dual-dimensional Uncertainty Framework for Streaming Active Learning in Industrial Defect Detection
为提高工业缺陷检测中深度学习模型的可靠性，本文提出基于模糊双维不确定性框架FuDU的流式主动学习方法，通过量化图像级和框级不确定性并融合专家知识进行自适应采样。
0 citationsRead paper
MAOL: Morphology-Aware Ordinal Learning for Fine-Grained Industrial Defect Severity Grading
本文提出MAOL框架，通过形态感知和序数学习方法解决工业缺陷精细分级问题，提高对不完美预测实例的鲁棒性。
0 citationsRead paper
AI worsens climate change, integrated assessment shows
研究提出一个整合AI对排放、产出和气候损害影响的框架，分析AI加剧气候变化的问题，区分ICT类和工业革命类AI前景，发现减缓措施与AI发展相辅相成。
0 citationsRead paper
Degenerate in Whose Frame? An Equivariance Condition for Degeneracy Detection in LiDAR Registration
This study addresses the inconsistency of LiDAR registration degeneracy detection labels caused by coordinate frame variations. We propose a generalized eigenvalue criterion based on an equivariant point displacement metric to resolve this issue. By employing adjoint reparameterization and generalized eigenvalue decomposition, we establish a novel frame-independent and scale-invariant paradigm for degeneracy determination that fundamentally eliminates reliance on specific reference frames. Experimental results demonstrate that the proposed criterion enables robust cross-sequence threshold transferability. Furthermore, our analysis reveals that correction magnitudes in 44.5%–69.5% of frame pairs are significantly affected by coordinate system choices, thereby validating the effectiveness and robustness of our approach in overcoming traditional frame-dependent limitations.
0 citationsRead paper
Recent publications
Latest Papers
SeaCausal-FL: Federated Fuzzy Causal Learning for Maritime IoT Fault Diagnosis and Counterfactual Reasoning
为解决海洋物联网中因数据分布和条件变化导致的故障诊断难题，提出SeaCausal-FL框架，结合联邦学习与模糊因果推理方法，提高了诊断准确性。
0 citationsRead paper
FuDU: A Fuzzy Dual-dimensional Uncertainty Framework for Streaming Active Learning in Industrial Defect Detection
为提高工业缺陷检测中深度学习模型的可靠性，本文提出基于模糊双维不确定性框架FuDU的流式主动学习方法，通过量化图像级和框级不确定性并融合专家知识进行自适应采样。
0 citationsRead paper
MAOL: Morphology-Aware Ordinal Learning for Fine-Grained Industrial Defect Severity Grading
本文提出MAOL框架，通过形态感知和序数学习方法解决工业缺陷精细分级问题，提高对不完美预测实例的鲁棒性。
0 citationsRead paper
AI worsens climate change, integrated assessment shows
研究提出一个整合AI对排放、产出和气候损害影响的框架，分析AI加剧气候变化的问题，区分ICT类和工业革命类AI前景，发现减缓措施与AI发展相辅相成。
0 citationsRead paper
Degenerate in Whose Frame? An Equivariance Condition for Degeneracy Detection in LiDAR Registration
This study addresses the inconsistency of LiDAR registration degeneracy detection labels caused by coordinate frame variations. We propose a generalized eigenvalue criterion based on an equivariant point displacement metric to resolve this issue. By employing adjoint reparameterization and generalized eigenvalue decomposition, we establish a novel frame-independent and scale-invariant paradigm for degeneracy determination that fundamentally eliminates reliance on specific reference frames. Experimental results demonstrate that the proposed criterion enables robust cross-sequence threshold transferability. Furthermore, our analysis reveals that correction magnitudes in 44.5%–69.5% of frame pairs are significantly affected by coordinate system choices, thereby validating the effectiveness and robustness of our approach in overcoming traditional frame-dependent limitations.
0 citationsRead paper