aiqbench analysis
Acacia represents a significant methodological shift in graph foundation models by demonstrating that complex capabilities like in-context learning and task generalization can emerge from training on raw web data without relying on pretrained large language models. This research validates the potential for graph-specific architectures to achieve autonomy similar to LLMs, offering a more efficient path toward versatile graph intelligence.
Training Graph Foundation Models on The Web Graph
arXiv:2609.30894v1 Announce Type: cross Abstract: We introduce Acacia, a graph foundation model, trained on the web graph. Acacia (i) supports arbitrary feature dimensionalities and semantics without additional training, (ii) supports a wide range of tasks, including node classification, link prediction, node clustering, and graph generation, without additional training, (iii) has in-context learning capabilities, and (iv) does not rely on pretrained LLMs. In particular, existing graph foundation models often require training additional classification heads or feature projectors to accommodate new graphs or new labels, whereas Acacia does not. Moreover, existing graph foundation models often gain their capabilities by being stitched together with pretrained LLMs, whereas Acacia is trained from scratch using only the Common Crawl web graph. This is also an important result because it provides evidence that graph models can acquire emergent capabilities from scratch like LLMs.
Abstract:We introduce Acacia, a graph foundation model, trained on the web graph. Acacia (i) supports arbitrary feature dimensionalities and semantics without additional training, (ii) supports a wide range of tasks, including node classification, link prediction, node clustering, and graph generation, without additional training, (iii) has in-context learning capabilities, and (iv) does not rely on pretrained LLMs. In particular, existing graph foundation models often require training additional classification heads or feature projectors to accommodate new graphs or new labels, whereas Acacia does not. Moreover, existing graph foundation models often gain their capabilities by being stitched together with pretrained LLMs, whereas Acacia is trained from scratch using only the Common Crawl web graph. This is also an important result because it provides evidence that graph models can acquire emergent capabilities from scratch like LLMs.
Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2609.30894 [cs.AI]
(or arXiv:2609.30894v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.30894 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.
Full text under the publisher's open licence · 229 words
Related coverage
GraphPFN: A Prior-Data Fitted Graph Foundation Model
arXiv · Aug 21, 2026
More Domains, No Gains? FLO: Fast-Slow Reciprocal Learning for Multi-Domain Graph Pre-training
arXiv · Oct 14, 2026
Handling Feature Heterogeneity with Learnable Graph Patches
arXiv · Jun 17, 2026
Presentation: Engineering AI for Creativity and Curiosity on Mobile
InfoQ · Jul 21, 2026