Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces HyPE, a novel framework for persona-grounded dialogue systems that encodes persona attributes using a hypergraph structure to capture high-order relations among these attributes. By analyzing persona texts as quadruples and employing Persistent Edge Embeddings within a HyperGCN neural network, HyPE generates a more nuanced persona summary vector that enhances response generation. The results show that HyPE significantly outperforms traditional sentence-level pooling methods across various model architectures, highlighting the effectiveness of structured hyperedge-level encoding in improving dialogue consistency.
Structured hypergraph encoding of persona attributes leads to superior dialogue consistency, outperforming traditional methods across multiple model architectures.
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.