analiza aiqbench
This arXiv preprint proposes a two-stage generative model that represents documents as typed graphs, using a Transformer for structure and a graph attention network for content, with explicit likelihoods that can flag forgeries. The idea of scoring document likelihood as an anomaly signal is practical, but the results are reported only on benchmark datasets (DocLayNet, FUNSD, SROIE), so real-world performance on forgery detection and editing remains to be validated.
FRAGMENT: Factorized Graph Representations for Document Generation and Editing via Entity-Aware Transformations
arXiv:2608.18679v1 Announce Type: cross Abstract: Structured documents such as invoices, forms, reports, and scientific articles derive meaning from the interplay between spatial layout, textual content, and logical structure. Generative models operating at the pixel or token level often struggle to capture these dependencies effectively. We explore FRAGMENT, a generative framework that represents a document as a typed relational graph and factorizes its distribution as p(structure, content) = p(structure) * p(content | structure). The framework consists of two stages. The first stage, the Architect, is a causally masked Transformer conditioned on document category that autoregressively generates the graph topology and typed spatial relations. The second stage, the Builder, is a GATv2-based graph attention network that enriches the graph with normalized bounding boxes, text, and visual style attributes. Both stages define explicit likelihood models, yielding a tractable document-level likelihood that serves as an anomaly score for forgery detection. For controlled editing, a prompt-conditioned extension injects instruction embeddings into the Builder through cross-attention, enabling semantic and entity-aware modifications. We describe training on DocLayNet and fine-tuning on FUNSD and SROIE. Experiments on DocLayNet, FUNSD, and SROIE evaluate FRAGMENT alongside representative autoregressive, layout-only, and graph-based baselines, providing an empirical analysis of the characteristics and trade-offs of the proposed factorized graph generation framework.
Abstract:Structured documents such as invoices, forms, reports, and scientific articles derive meaning from the interplay between spatial layout, textual content, and logical structure. Generative models operating at the pixel or token level often struggle to capture these dependencies effectively. We explore FRAGMENT, a generative framework that represents a document as a typed relational graph and factorizes its distribution as p(structure, content) = p(structure) * p(content | structure). The framework consists of two stages. The first stage, the Architect, is a causally masked Transformer conditioned on document category that autoregressively generates the graph topology and typed spatial relations. The second stage, the Builder, is a GATv2-based graph attention network that enriches the graph with normalized bounding boxes, text, and visual style attributes. Both stages define explicit likelihood models, yielding a tractable document-level likelihood that serves as an anomaly score for forgery detection. For controlled editing, a prompt-conditioned extension injects instruction embeddings into the Builder through cross-attention, enabling semantic and entity-aware modifications. We describe training on DocLayNet and fine-tuning on FUNSD and SROIE. Experiments on DocLayNet, FUNSD, and SROIE evaluate FRAGMENT alongside representative autoregressive, layout-only, and graph-based baselines, providing an empirical analysis of the characteristics and trade-offs of the proposed factorized graph generation framework.
Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
arXiv:2608.18679 [cs.CV]
(or arXiv:2608.18679v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2608.18679 Focus to learn more arXiv-issued DOI via DataCite
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 · 301 cuvinte
Subiecte conexe
Understanding Rate-Distortion Performance in Distributed Transformer Inference
arXiv · 4 aug. 2026
Generalized belief states in transformers
LessWrong · 31 oct. 2026
Scaffold Then Internalize: Representation Injection for Diffusion Transformers
arXiv · 30 sept. 2026
Global Divergence, Local Convergence: Representation Geometry in SSMs and Transformers
arXiv · 9 sept. 2026