HyPE: Category-Aware Hypergraph Encoding with Persistent Edge Embeddings for Persona-Grounded Dialogue

2026-06-11

AI Summary

This work addresses the limitation of existing dialogue systems that model persona descriptions as flat sentences, which fails to capture higher-order semantic relationships such as shared thematic categories. To overcome this, the authors propose HyPE, a novel framework that introduces a category-aware hypergraph structure. Specifically, persona attributes are parsed into quadruples—(core, expression, emotion, category)—and categories serve as hyperedges to construct the hypergraph. The framework further incorporates lightweight Persistent Edge Embeddings (PEE) as learnable priors during message passing. Experiments on PersonaChat demonstrate that HyPE, when integrated with diverse language models including GPT-2, LLaMA-3.2-3B, and Qwen2.5-3B, consistently outperforms sentence-level pooling baselines, confirming its effectiveness and transferability across different model scales.

Abstract

Persona-grounded dialogue systems aim to produce responses consistent with a speaker's persona, yet existing methods treat personas as a flat set of sentences and fail to model the high-order relations among persona attributes-e.g., that several persona sentences share a topical category. We propose HyPE (Hypergraph Persona Encoder), a framework that (i) analyzes each persona-bearing text as a (Core, Expression, Sentiment, Category) quadruple, and (ii) organizes persona elements into a hypergraph whose hyperedges are induced by shared category labels. An HyperGCN hypergraph neural network propagates this structure into a persona summary vector and a soft-memory bank that condition the response generator. We further propose Persistent Edge Embeddings (PEE), lightweight per-category learnable priors fused into the HyperGCN message-passing step. On PersonaChat under greedy decoding, HyPE consistently outperforms sentence-level pooling baselines across GPT-2, LLaMA-3.2-3B, and Qwen2.5-3B backbones by demonstrating that structured hyperedge-level persona encoding provides a transferable advantage across model scales.

Problem

• persona-grounded dialogue
• hypergraph
• high-order relations
• category-aware
• persona encoding

Innovation

• Hypergraph Neural Network
• Persona-Grounded Dialogue
• Persistent Edge Embeddings
• High-Order Relations
• Category-Aware Encoding