Factual and Edit-Sensitive Graph-to-Sequence Generation via Graph-Aware Adaptive Noising

TL;DR

This paper introduces DLM4G, a non-autoregressive diffusion model for graph-to-sequence generation that enhances factual accuracy and edit sensitivity by adaptive noising and iterative refinement.

Contribution

It proposes a novel diffusion framework with adaptive noising for improved graph-to-sequence generation, outperforming autoregressive baselines and demonstrating broad applicability.

Findings

DLM4G outperforms diffusion baselines on multiple datasets and metrics.

It exceeds larger autoregressive models in factual grounding and edit sensitivity.

Shows effectiveness in scientific graph-to-sequence tasks like molecule captioning.

Abstract

Fine-tuned autoregressive models for graph-to-sequence generation (G2S) often struggle with factual grounding and edit sensitivity. To tackle these issues, we propose a non-autoregressive diffusion framework that generates text by iterative refinement conditioned on an input graph, named as Diffusion Language Model for Graphs (DLM4G). By aligning graph components (entities/relations) with their corresponding sequence tokens, DLM4G employs an adaptive noising strategy. The proposed strategy uses per-token denoising error as a signal to adaptively modulate noise on entity and relation tokens, improving preservation of graph structure and enabling localized updates under graph edits. Evaluated on three datasets, DLM4G consistently outperforms competitive G2S diffusion baselines trained on identical splits across both surface-form and embedding-based metrics. DLM4G further exceeds fine-tuned autoregressive baselines up to 12x larger (e.g., T5-Large) and is competitive with zero-shot LLM transfer baselines up to 127x larger. Relative to the strongest fine-tuned PLM baseline, DLM4G improves factual grounding (FGT@0.5) by +5.16% and edit sensitivity (ESR) by +7.9%; compared to the best diffusion baseline, it yields gains of +3.75% in FGT@0.5 and +23.6% in ESR. We additionally demonstrate applicability beyond textual graphs through experiments on molecule captioning, indicating the method's generality for scientific G2S generation.