Tag
This preprint proposes a protection-oriented assessment framework for ML-based fault detectors in power systems, evaluating three models and a sequence-current reference on the PROTECT-90 dataset of 9,022 simulated fault episodes. Despite all models achieving AUC–ROC above 0.999, missed-trip rates varied by over two orders of magnitude at a fixed security budget, showing that standard metrics hide critical protection trade-offs.
该论文确定了在独立非同分布噪声和两两非负相关噪声下,选择最大噪声分数所产生的“赢家偏差”的尖锐常数界,并将全部定理在 Lean 4 中形式化验证,同时在 448 个模型选择流程上验证了结论的实际意义。
该研究评估了一种用于慢性踝关节不稳的自适应步态生物反馈系统,采用时间卷积分类器和留一被试交叉验证(AUROC 0.948),并结合被试特异性模型更新;结果显示自适应组在额状面踝关节角度上较对照组有显著改善。
Researchers propose a calibrated hybrid ensemble combining five deep learning and three classical ML models via a logistic regression stacking meta-learner for EEG-based epileptic seizure forecasting. Evaluated on CHB-MIT with Leave-One-Patient-Out cross-validation, the model reaches 74.2% seizure-level sensitivity at 1.24 false alarms per hour with ~16.9 minutes of warning time.
A non-coder built a site where AI agents from different vendors independently check each other's work, leading to the discovery and fixing of real governance loopholes on the first day.
This paper presents an evaluation methodology for LLM security detectors that addresses systematic weaknesses like per-dataset threshold tuning and undisclosed operating points. The framework uses cross-validation across 16 benchmarks, selects a single global operating point, and includes multiple diagnostics for generalization.
This paper compares cross-validation ensembles to deep ensembles for uncertainty estimation in medical image segmentation. Deep ensembles outperform cross-validation ensembles in calibration and failure detection, while cross-validation ensembles better approximate inter-rater variability.