Long Time No See! Open-Domain Conversation with Long-Term Persona - Memory - 2022 Findings of the Association for Computational Linguistics (ACL 2022)

Aug 27, 2026

Most open-domain dialogue models struggle in long-term human-bot conversations because they cannot effectively understand and memorize dialogue history. To address this gap, the paper introduces the Long-term Memory Conversation (LeMon) task, builds the DuLeMon dataset, and proposes PLATO-LTM, a generation framework with a long-term memory mechanism. The mechanism extracts and continuously updates mutual persona memory for both user and bot in real time, improving long-dialogue consistency and engagement.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Introduction

Long-term persona ability and its challenges

Related work: persona dialogue datasets and models

Proposed LeMon task and DuLeMon dataset

PLATO-LTM framework with long-term memory

Long Time No See! Open-Domain Conversation with Long-Term Persona Memory Xinchao Xu1∗, Zhibin Gou1,2∗, Wenquan Wu1 , Zheng-Yu Niu1 , Hua Wu1 , Haifeng Wang1 and Shihang Wang3 1Baidu Inc., China 2 School of Computer Science, Beijing University of Posts and Telecommunications 3 Columbia University {xinchaoxu,wuwenquan01,niuzhengyu,wu_hua,[wanghaifeng}@baidu.com](wanghaifeng}@baidu.com)[ ](wanghaifeng}@baidu.com)[zebgou@gmail.com](zebgou@gmail.com), [sw3275@columbia.edu](sw3275@columbia.edu) Abstract Most of the open-domain dialogue models tend to perform poorly in the setting of long-term human-bot conversations. The possible reason is that they lack the capability of understanding and memorizing long-term dialogue history information. To address this issue, we present a novel task of Long-term Memory Conversation (LeMon) and then build a new dialogue dataset DuLeMon and a dialogue generation framework PLATO-LTM with a Long-Term Memory (LTM) mechanism. This LTM mechanism enables our system to accurately extract and continuously update long-term persona memory without requiring multiple-session dialogue datasets for model training. To our knowledge, this is the first attempt to conduct real-time dynamic management of persona information of both parties, including the user and the bot. Results on DuLeMon indicate that PLATO-LTM can significantly outperform baselines in terms of long-term dialogue consistency, leading to better dialogue engagingness 1. 1 Introduction Persona is crucial for open-domain dialogue systems to establish long-term intimacy with users (Huang et al., 2020) . Existing persona dialogue datasets such as PersonaChat (Zhang et al., 2018 ; Dinan et al., 2019) and models (Li et al., 2016a ; Zhang et al., 2017 ; Qian et al., 2018) have greatly facilitated the chatbot with configurable and persistent personalities. Nevertheless, current open-domain dialogue systems still cannot build a long-term connection with humans. The possible reason is that they lack the capability of understanding and memorizing longterm dialogue history information, which we called ∗ Equal contribution. The work was done when Zhibin Gou and Shihang Wang were doing internship at Baidu. 1Our data and codes are released at [https:](https:)//[github.com/PaddlePaddle/Research/tree/](github.com/PaddlePaddle/Research/tree/)[ ](github.com/PaddlePaddle/Research/tree/)master/NLP/ACL2022-DuLeMon User Chatbot Trigger Trigger Figure 1: A sample of long-term conversation with memory. At first, the chat partner is not familiar with each other, so the goal is to get to know each other; Then, after multiple sessions, the chatbot already has a certain understanding and memory of the user’s persona and its own persona, making the deep chat possible. long-term persona ability. Remembering and actively utilizing the user’s persona increases engagingness and contributes to long-term friendships between chatbot and user (Campos et al., 2018) . Without this ability, the current state-of-the-art models, such as Meena (Adiwardana et al., 2020), Blender (Roller et al., 2021), and PLATO (Bao et al., 2020), tend to talk to people like strangers in long-term conversations. Despite the importance and challenge of utilizing long-term persona in open-domain dialogue, as far as we know, the long-term persona ability of large-scale models is less studied due to a lack of both task design and corresponding dataset. Previous long-term persona dialogue systems (Kim et al., 2014 ; Bang et al., 2015) are mainly rule-based systems without large-scale pre-training models, in which researchers proposed various episodic memory architectures to extract, store and manage rel- 2639 Findings of the Association for Computational Linguistics: ACL 2022 , pages 2639-2650 May 22-27, 2022 􀀍c2022 Association for Computational Linguistics evant facts in prior interactions for use in future dialogs (Campos et al., 2018) . In addition, existing person

FAQs

为什么现有开放域对话模型在长期人机对话中表现较差？

主要原因是模型缺乏理解并记忆长期对话历史信息的能力，无法稳定维持用户与机器之间的长期关联。

论文提出了哪些关键方法来解决长期人格记忆问题？

提出了Long-term Memory Conversation（LeMon）任务，构建DuLeMon数据集，并开发带长时记忆机制的PLATO-LTM对话生成框架。

PLATO-LTM如何在对话过程中管理并更新长期人格信息？

PLATO-LTM能够在实时对话中抽取双方人格信息，并持续写入与更新长期人格记忆，从而支持后续更一致、更具参与感的对话。

Aug 27, 2026

Upload by วิน