9 papers
Same Loss, Different Gradients
Ningkang Peng, Xiaoqian Peng, Yifan He +4
Differentiable learning typically assumes that the scalar objective evaluated in the forward pass and the gradient supplied to the optimizer in the backward pass describe the same…
Gradient Mirage: Trainable yet Label-Unidentifiable Gradients in Large Language Model Split Learning
Shiyu Miao, Yunlong Mao, Zirui Huang +5
Gradient matching attacks (GMAs) in LLM split learning (SL) rely on a critical yet underexplored assumption: the gradient exposed at the split interface is a faithful derivative of…
Early High-Frequency Injection for Geometry-Sensitive OOD Detection
Chuanjie Cheng, Ningkang Peng, Chenxi Liu +3
Post-hoc OOD detectors score logits or features after training, so their success depends on the geometry already encoded in the representation. We revisit this assumption through a…
GAMR: Geometric-Aware Manifold Regularization with Virtual Outlier Synthesis for Learning with Noisy Labels
Ningkang Peng, Jingyang Mao, Xiaoqian Peng +4
Deep neural networks (DNNs) experience significant performance degradation when processing noisy labels, primarily due to overfitting on mislabeled data. Current mainstream approac…
Holistic Reliability Propagation: Decoupling Annotation and Prediction for Robust Noisy-Label
Jingyang Mao, Ningkang Peng, Yanhui Gu
Learning with noisy labels in multimedia classification often combines external annotations and model predictions into a single reliability weight, even though the two sources can…
Is Complex Training Necessary for Long-Tailed OOD Detection? A Re-think from Feature Geometry
Ningkang Peng, Xuanming Chen, Yanhui Gu
Long-tailed out-of-distribution (LT-OOD) detection is often addressed with specialized training, including auxiliary out-of-distribution (OOD) data, abstention heads, contrastive o…
When Accuracy Is Not Enough: Uncertainty Collapse between Noisy Label Learning and Out-of-Distribution Detection
Ningkang Peng, Jingyang Mao, Runhan Zhou +2
Learning with noisy labels (LNL) is typically benchmarked by closed-set classification accuracy, yet deployment often requires classifiers to reject out-of-distribution (OOD) input…
Radial-Angular Geometry for Reliable Update Diagnosis in Noisy-Label Learning
Ningkang Peng, Jingyang Mao, Xiaoqian Peng +2
Noisy-label methods often estimate sample reliability from forward-space signals such as loss, confidence, or entropy. These signals indicate whether a sample is difficult to predi…
HamBR: Active Decision Boundary Restoration Based on Hamiltonian Dynamics for Learning with Noisy Labels
Ningkang Peng, Jingyang Mao, Qianfeng Yu +3
In large-scale visual recognition and data mining tasks, the presence of noisy labels severely undermines the generalization capability of deep neural networks (DNNs). Prevalent sa…