HEISIR: Hierarchical Expansion of Inverted Semantic Indexing for Training-free Retrieval of Conversational Data using LLMs

2025-03-06

AI Summary

Existing conversational retrieval methods rely heavily on large-scale annotation and fine-tuning, compromising the trade-off between semantic modeling accuracy and deployment efficiency. To address this, we propose a zero-shot, zero-annotation hierarchical inverted semantic indexing framework. It extracts SVOA (Subject–Verb–Object–Adjunct) quatruples via dependency parsing and semantic role labeling, and explicitly models temporal, intentional, and contextual adjunct semantics through an Adjunct Augmentation mechanism. Furthermore, we design a plug-and-play retrieval architecture compatible with embeddings from multiple large language models (LLMs), requiring no training. Experiments on multi-turn conversational retrieval show an average 12.7% improvement in Recall@10 and a 68% reduction in latency. The framework has been successfully deployed in three industrial dialogue systems, supporting intent recognition and topic evolution analysis.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Conversational AI/Dialog Systems
• Computer Vision: Image and Video Retrieval

Application Category

• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

The growth of conversational AI services has increased demand for effective information retrieval from dialogue data. However, existing methods often face challenges in capturing semantic intent or require extensive labeling and fine-tuning. This paper introduces HEISIR (Hierarchical Expansion of Inverted Semantic Indexing for Retrieval), a novel framework that enhances semantic understanding in conversational data retrieval through optimized data ingestion, eliminating the need for resource-intensive labeling or model adaptation. HEISIR implements a two-step process: (1) Hierarchical Triplets Formulation and (2) Adjunct Augmentation, creating semantic indices consisting of Subject-Verb-Object-Adjunct (SVOA) quadruplets. This structured representation effectively captures the underlying semantic information from dialogue content. HEISIR achieves high retrieval performance while maintaining low latency during the actual retrieval process. Our experimental results demonstrate that HEISIR outperforms fine-tuned models across various embedding types and language models. Beyond improving retrieval capabilities, HEISIR also offers opportunities for intent and topic analysis in conversational data, providing a versatile solution for dialogue systems.

Problem

• Enhances semantic understanding in conversational data retrieval
• Eliminates need for resource-intensive labeling or model adaptation
• Improves retrieval performance and reduces latency in dialogue systems

Innovation

• Hierarchical Triplets Formulation for semantic indexing
• Adjunct Augmentation enhances SVOA quadruplets representation
• Optimized data ingestion eliminates need for labeling