PK-ICR: Persona-Knowledge Interactive Context Retrieval for Grounded Dialogue

TL;DR

This paper introduces PK-ICR, a novel method for jointly retrieving persona and knowledge contexts in complex dialogues, improving grounding efficiency with less computational power and a new null-positive rank test for better evaluation.

Contribution

The paper proposes a dual context identification task and a neural QA retrieval method that handles multiple contexts simultaneously, enhancing grounded dialogue systems.

Findings

Effective joint retrieval of persona and knowledge contexts.

Reduces computational requirements compared to existing methods.

Introduces a null-positive rank test for evaluating retrieval robustness.

Abstract

Identifying relevant persona or knowledge for conversational systems is critical to grounded dialogue response generation. However, each grounding has been mostly researched in isolation with more practical multi-context dialogue tasks introduced in recent works. We define Persona and Knowledge Dual Context Identification as the task to identify persona and knowledge jointly for a given dialogue, which could be of elevated importance in complex multi-context dialogue settings. We develop a novel grounding retrieval method that utilizes all contexts of dialogue simultaneously. Our method requires less computational power via utilizing neural QA retrieval models. We further introduce our novel null-positive rank test which measures ranking performance on semantically dissimilar samples (i.e. hard negatives) in relation to data augmentation.