跳到正文
arXiv cs.AI· Xi Ding, Lei Wang, Piotr Koniusz, Yongsheng Gao·· 4 小时前AI 评分32

论文《Graph Your Own Prompt》提出图一致性正则化框架 GCR

Graph Your Own Prompt

AI 导读

研究者提出图一致性正则化框架 GCR,通过将模型预测派生的关系图结构注入学习过程,实现"自我提示"式的表示优化。该框架在任意深度插入无参数的图一致性层(GCL),构建批次级特征相似度图并与类感知掩码预测图对齐,且层重要性由图差异幅度自适应学习。GCR 与具体模型无关、轻量且不改架构,实验显示其提升特征结构、类内凝聚性与泛化能力,论文已被 NeurIPS 2025 接收。

正文 · 原文

View PDF HTML (experimental)

Abstract:We propose Graph Consistency Regularization (GCR), a novel framework that injects relational graph structures, derived from model predictions, into the learning process to promote class-aware, semantically meaningful feature representations. Functioning as a form of self-prompting, GCR enables the model to refine its internal structure using its own outputs. While deep networks learn rich representations, these often capture noisy inter-class similarities that contradict the model's predicted semantics. GCR addresses this issue by introducing parameter-free Graph Consistency Layers (GCLs) at arbitrary depths. Each GCL builds a batch-level feature similarity graph and aligns it with a global, class-aware masked prediction graph, derived by modulating softmax prediction similarities with intra-class indicators. This alignment enforces that feature-level relationships reflect class-consistent prediction behavior, acting as a semantic regularizer throughout the network. Unlike prior work, GCR introduces a multi-layer, cross-space graph alignment mechanism with adaptive weighting, where layer importance is learned from graph discrepancy magnitudes. This allows the model to prioritize semantically reliable layers and suppress noisy ones, enhancing feature quality without modifying the architecture or training procedure. GCR is model-agnostic, lightweight, and improves semantic structure across various networks and datasets. Experiments show that GCR promotes cleaner feature structure, stronger intra-class cohesion, and improved generalization, offering a new perspective on learning from prediction structure. [Project website](this https URL) [Code](this https URL)
Comments: NeurIPS 2025 (Author-corrected version: Revised by the original authors, October 2026)
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2509.23373 [cs.LG]
  (or arXiv:2509.23373v4 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2509.23373

arXiv-issued DOI via DataCite

Submission history

From: Lei Wang [view email]
[v1] Sat, 27 Sep 2025 15:45:07 UTC (32,003 KB)
[v2] Sun, 12 Oct 2025 15:43:09 UTC (32,003 KB)
[v3] Mon, 28 Sep 2026 14:00:50 UTC (1 KB) (withdrawn)
[v4] Thu, 12 Nov 2026 22:34:43 UTC (31,995 KB)

来源:arXiv cs.AI · arxiv.org