paper-with-me

홈 › Papers

FragileFlow: Spectral Control of Correct-but-Fragile Predictions for Foundation Model Robustness

2026-05-09 · Zhuoyun Li, Boxuan Wang, Jinwei Hu, Xiaowei Huang, Yi Dong arxiv

Robust adaptation of LLMs and VLMs is often evaluated by average accuracy or average consistency under perturbations. However, these averages can hide a structured failure mode: a prediction may remain correct while probability mass already flows from particular true classes toward systematic wrong competitors near the decision boundary. In this paper, we formalize this phenomenon as margin-aware error flow and introduce FragileFlow, a plug-in regularizer that uses a calibrated margin buffer to identify correct-but-fragile predictions and organize their off-class probability mass into a class-wise vulnerable-risk matrix. Theoretically, we provide the first PAC-Bayes upper bound for this margin-aware error-flow object, showing how empirical spectral control yields a conservative route to deterministic worst-class robustness under a stability condition. Experiments on multiple-choice LLM benchmarks and few-shot CLIP adaptation show that FragileFlow consistently improves the proposed theory-facing risk measures over matched baselines, yields perturbed worst-class accuracy gains in most settings, and preserves clean accuracy across comparisons.

📄 PDF Abstract BibTeX arXiv:2605.08896

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Spectral-Spatial Synergistic Guided Network for Hyperspectral Salient Object Detection

2026-07-23 · Yanyan Peng, Tingfa Xu, Yao Xiao, Peifu Liu 외 arxiv

Hyperspectral salient object detection aims to identify visually salient regions from hyperspectral images. Existing methods often fail because they fundamentally misunderstand the data, confusing incidental spectral var…

Computational EfficiencySalient Object Detection

Graph Neural Network Explanations are Fragile

2024-06-05 · Jiate Li, Meng Pang, Yun Dong, Jinyuan Jia 외

Explainable Graph Neural Network (GNN) has emerged recently to foster the trust of using GNNs. Existing GNN explainers are developed from various perspectives to enhance the explanation performance. We take the first ste…

Adversarial AttackGraph Neural Network

Three-Pronged Spectral Control for Federated Parameter Efficient Fine Tuning

2026-07-23 · Shiva Raj Pokhrel, Dipsan Bhattarai, Anwar Walid arxiv

Federated parameter-efficient fine-tuning (PEFT) enables communication-efficient adaptation of large pretrained models on decentralized edge data, but it remains fragile under non-IID client heterogeneity. In low-rank ad…

parameter-efficient fine-tuning

Quasi-Framelets: Robust Graph Neural Networks via Adaptive Framelet Convolution

2022-01-11 · Mengxi Yang, Dai Shi, Xuebin Zheng, Jie Yin 외

This paper aims to provide a novel design of a multiscale framelet convolution for spectral graph neural networks (GNNs). While current spectral methods excel in various graph learning tasks, they often lack the flexibil…

Graph LearningGraph Neural NetworkNode Classification

Breaking the Chain: A Causal Analysis of LLM Faithfulness to Intermediate Structures

2026-03-17 · Oleg Somov, Mikhail Chaichuk, Gleb Ershov, Karim Vafin 외 arxiv

In schema-guided reasoning (SGR) pipelines, LLMs produce explicit intermediate structures -- rubrics, checklists, or verification queries -- before committing to a final decision. SGR is increasingly adopted because it p…