Enabling Real-Time Conversations with Minimal Training Costs

TL;DR

This paper introduces a duplex decoding method that enables real-time conversational interactions with large language models, requiring minimal additional training and computational resources, thus improving naturalness and human-likeness.

Contribution

The paper proposes a novel duplex decoding approach that enhances LLMs with real-time conversational ability using minimal training, addressing computational overhead issues.

Findings

Significantly improves interaction naturalness and human-likeness

Requires minimal additional training and computational resources

Effective in enabling real-time conversations

Abstract

Large language models (LLMs) have demonstrated the ability to improve human efficiency through conversational interactions. Conventional LLM-powered dialogue systems, operating on a turn-based paradigm, preclude real-time interaction during response generation. To address this limitation, researchers have proposed duplex models. These models can dynamically adapt to user input, facilitating real-time interactive feedback. However, these methods typically require substantial computational resources to acquire the ability. To reduce overhead, this paper presents a new duplex decoding approach that enhances LLMs with duplex ability, requiring minimal additional training. Specifically, our method employs parallel decoding of queries and responses in conversations, effectively implementing a channel-division-multiplexing decoding strategy. Experimental results indicate that our proposed method significantly enhances the naturalness and human-likeness of user-AI interactions with minimal training costs.