Hello Again! LLM-powered Personalized Agent for Long-term Dialogue

Open-domain dialogue systems have seen remarkable advancements with the development of large language models (LLMs). Nonetheless, most existing dialogue systems predominantly focus on brief single-session interactions, neglecting the real-world demands for long-term companionship and personalized interactions with chatbots. Crucial to addressing this real-world need are event summary and persona management, which enable reasoning for appropriate long-term dialogue responses. Recent progress in the human-like cognitive and reasoning capabilities of LLMs suggests that LLM-based agents could significantly enhance automated perception, decision-making, and problem-solving. In response to this potential, we introduce a model-agnostic framework, the Long-term Dialogue Agent (LD-Agent), which incorporates three independently tunable modules dedicated to event perception, persona extraction, and response generation. For the event memory module, long and short-term memory banks are employed to separately focus on historical and ongoing sessions, while a topic-based retrieval mechanism is introduced to enhance the accuracy of memory retrieval. Furthermore, the persona module conducts dynamic persona modeling for both users and agents. The integration of retrieved memories and extracted personas is subsequently fed into the generator to induce appropriate responses. The effectiveness, generality, and cross-domain capabilities of LD-Agent are empirically demonstrated across various illustrative benchmarks, models, and tasks.

Motivated by real-life demands, the core challenge of open-domain dialogue systems is to simultaneously maintain long-term event memory and preserve persona consistency [9–11, 3]. Existing research often addresses these aspects separately—focusing either on event memory or persona extraction—thereby hindering long-term consistency. Current strategies for event memory typically involve constructing a memory bank that stores historical event summaries, complemented by retrieval-augmented approaches to access relevant information for response generation [12, 13]. Studies on persona-based dialogue rang from unidirectional user modeling [14] to bidirectional agent-user modeling [15, 16, 3], enhancing personalized chat abilities by leveraging profile information. Worse still, the aforementioned methods are highly dependent on specific model architectures, making them challenging to adapt to other models.