Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Clustering as Reasoning: A $k$-Means Interpretation of Chain-of-Thought Graph Learning

About

Chain-of-Thought (CoT) prompting has shown promise in enhancing the reasoning capabilities of large language models (LLMs) on text-attributed graphs (TAGs). This work reframes CoT-based graph learning through the principle of clustering as reasoning, offering a $k$-means interpretation of how iterative reasoning operates over graph-structured data. We observe that existing graph CoT methods rely on disjoint architectures and fixed graph representations, limiting step-by-step semantic-topological interaction and interpretability. To overcome this limitation, we propose a unified framework named KCoT that integrates CoT reasoning with graph representation learning. Our key theoretical result reveals a formal mathematical correspondence between a Transformer block and the $k$-means algorithm, allowing reasoning to be interpreted as iterative assignment and update steps. Based on this insight, we introduce a Semantic Discriminating Prompt that explicitly formulates these steps as structured CoT reasoning, together with a structure-grounded alignment strategy to fuse topological priors with evolving thought-conditioned representations. Experiments on standard benchmarks demonstrate consistent improvements over state-of-the-art methods, validating clustering as a principled mechanism for CoT-based graph learning.

Xuanting Xie, Zhaochen Guo, Bingheng Li, Xingtong Yu, Zhifei Liao, Zhao Kang, Yuan Fang• 2026

Related benchmarks

TaskDatasetResultRank
Node ClassificationCora
Accuracy91.64
583
Node ClassificationPubmed
Accuracy96.79
363
Node ClassificationarXiv
Accuracy79.26
254
Node ClassificationProducts
Accuracy86.8
85
Link PredictionPubmed
Accuracy95.97
33
Link PredictionarXiv
Accuracy95.72
29
Link PredictionProducts
Accuracy97.3
29
Showing 7 of 7 rows

Other info

Follow for update