analiza aiqbench
This research proposes a novel framework for dynamic edge classification in text-attributed graphs by allowing language model embeddings to co-evolve with network topology, rather than keeping them static. By integrating a lightweight Transformer adapter with a maintenance objective to anchor semantic structure, the method improves linear separability and classification performance without incurring prohibitive computational costs. The study demonstrates that adaptive semantic representations significantly outperform frozen-text baselines across diverse web-related domains like review networks and news graphs.
An Analysis of Semantic and Spatio-Temporal Co-evolution for Edge Classification on Dynamic Text-Attributed Graphs
arXiv:2610.25906v1 Announce Type: new Abstract: Dynamic edge classification on text-attributed Web graphs, such as review networks, community forums, and news-derived knowledge graphs, requires high quality, adaptive semantic representations in the context of network evolution. Existing dynamic graph architectures typically initialize text feature vectors with pretrained language models and hold them static during training to preserve their semantic structure, forgoing the potential benefits of allowing them to co-evolve with network topology. In this work, we investigate the effects of enabling such co-evolution with a unified text-graph alignment framework that integrates a lightweight Transformer adapter to update token-level text representations via spatio-temporal gradients from a contrastive pretraining objective. To prevent detrimental semantic distortion, a language model maintenance objective anchors the evolving text embeddings to their original semantic structure. Evaluations on the Dynamic Text-Attributed Graph Benchmark demonstrate that this anchored text-graph co-evolution consistently improves upon the edge classification performance of frozen-text baselines across diverse Web-related domains. Representational analyses reveal that the framework reduces anisotropy, improving the linear separability of text embeddings, and effectively encodes structural and temporal information into the semantic space. A computational efficiency analysis confirms that this beneficial semantic and spatio-temporal co-evolution is achieved without incurring the prohibitive computational overhead typical of end-to-end joint training.
Abstract:Dynamic edge classification on text-attributed Web graphs, such as review networks, community forums, and news-derived knowledge graphs, requires high quality, adaptive semantic representations in the context of network evolution. Existing dynamic graph architectures typically initialize text feature vectors with pretrained language models and hold them static during training to preserve their semantic structure, forgoing the potential benefits of allowing them to co-evolve with network topology. In this work, we investigate the effects of enabling such co-evolution with a unified text-graph alignment framework that integrates a lightweight Transformer adapter to update token-level text representations via spatio-temporal gradients from a contrastive pretraining objective. To prevent detrimental semantic distortion, a language model maintenance objective anchors the evolving text embeddings to their original semantic structure. Evaluations on the Dynamic Text-Attributed Graph Benchmark demonstrate that this anchored text-graph co-evolution consistently improves upon the edge classification performance of frozen-text baselines across diverse Web-related domains. Representational analyses reveal that the framework reduces anisotropy, improving the linear separability of text embeddings, and effectively encodes structural and temporal information into the semantic space. A computational efficiency analysis confirms that this beneficial semantic and spatio-temporal co-evolution is achieved without incurring the prohibitive computational overhead typical of end-to-end joint training.
Machine Learning (cs.LG)
arXiv:2610.25906 [cs.LG]
(or arXiv:2610.25906v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.25906 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 293 cuvinte
Subiecte conexe
Transfer Learning for Edge Classification on Dynamic Text-Attributed Graphs
arXiv · 30 sept. 2026
TTGBench: Benchmarking Topological Evolution and Semantic Drift in Text-attributed Temporal Graphs
arXiv · 9 sept. 2026
CausalPOI: Spatio-Temporal Graph-Based Causal Modeling for Cold-Start POI Check-in Forecasting
arXiv · 6 iun. 2026
HERMES: Heterogeneous Edge-Relational Multi-Head Embedded SSM Attention for Traffic Conflict Prediction at Signalized Intersections
arXiv · 24 iul. 2026