analiza aiqbench
This research addresses a key limitation in Neural Relational Inference (NRI) where traditional uniform priors fail to capture real-world structural dependencies. By introducing a diffusion-parameterized adaptive prior, the proposed Diff-prior framework calibrates latent graph distributions to produce more decisive edge posteriors during inference. The approach demonstrates improved performance across standard benchmarks and multiple NRI architectures, offering a more robust paradigm for structure discovery tasks.
From Uniform to Learned Graph Priors: Diffusion for Structure Discovery
arXiv:2606.11831v1 Announce Type: new Abstract: Neural relational inference (NRI) methods discover interaction graphs from trajectories through variational reasoning on discrete potential edges. However, these methods typically rely on oversimplified, factorized graph priors. Such priors, typically nearing uniform distributions, treat edges as independent entities. This systemic misalignment does not match the real-world systems and yields diffuse and indecisive edge posteriors limiting the reliability of structural discovery. To address this, we propose \textit{Diff-prior}, a diffusion-parameterized adaptive prior used to calibrate latent graph distribution rather than generate graphs. Our core insight is to reframe prior integration as a learnable denoising-style calibration that organizes scattered, uncertain edge posteriors into a more reliable overall structure which can be trained by the diffusion model. Diff-prior learns an adaptive structure prior that performs structured calibration on the edge posteriors during inference, guiding it towards a distribution closer to the underlying structure. The diff-prior operates before structural sampling and acts as a denoising calibrator directly on the encoder edge distribution, which provides a generic training paradigm over structured variables. Experiments on standard benchmarks validated our framework, and the results indicate that Diff-prior improves the performance of structure inference and generates more decisive edge posteriors across multiple NRI-family architectures. The code is available on https://github.com/Hardy158118/Diffprior.
Abstract:Neural relational inference (NRI) methods discover interaction graphs from trajectories through variational reasoning on discrete potential edges. However, these methods typically rely on oversimplified, factorized graph priors. Such priors, typically nearing uniform distributions, treat edges as independent entities. This systemic misalignment does not match the real-world systems and yields diffuse and indecisive edge posteriors limiting the reliability of structural discovery. To address this, we propose \textit{Diff-prior}, a diffusion-parameterized adaptive prior used to calibrate latent graph distribution rather than generate graphs. Our core insight is to reframe prior integration as a learnable denoising-style calibration that organizes scattered, uncertain edge posteriors into a more reliable overall structure which can be trained by the diffusion model. Diff-prior learns an adaptive structure prior that performs structured calibration on the edge posteriors during inference, guiding it towards a distribution closer to the underlying structure. The diff-prior operates before structural sampling and acts as a denoising calibrator directly on the encoder edge distribution, which provides a generic training paradigm over structured variables. Experiments on standard benchmarks validated our framework, and the results indicate that Diff-prior improves the performance of structure inference and generates more decisive edge posteriors across multiple NRI-family architectures. The code is available on this https URL.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2606.11831 [cs.LG]
(or arXiv:2606.11831v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.11831 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
https://doi.org/10.1145/3770855.3817940 Focus to learn more DOI(s) linking to related resources
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 · 310 cuvinte
Subiecte conexe
Efficient Graph Generation via Direct Prediction and Flow Matching
arXiv · 6 oct. 2026
Generating Graph-Like Logical Rules for Knowledge Graph Reasoning via Diffusion Models
arXiv · 6 iun. 2026
The payload: the semantic layer, memory, and the loop that fills it
Neo4j · 10 aug. 2026
Minimax Optimality of Score-Entropy Discrete Diffusion
arXiv · 24 aug. 2026