{"title":"通过约束满足评估面向任务的对话一致性","authors":"Tiziano Labruna, Bernardo Magnini","doi":"arxiv-2407.11857","DOIUrl":null,"url":null,"abstract":"Task-oriented dialogues must maintain consistency both within the dialogue\nitself, ensuring logical coherence across turns, and with the conversational\ndomain, accurately reflecting external knowledge. We propose to conceptualize\ndialogue consistency as a Constraint Satisfaction Problem (CSP), wherein\nvariables represent segments of the dialogue referencing the conversational\ndomain, and constraints among variables reflect dialogue properties, including\nlinguistic, conversational, and domain-based aspects. To demonstrate the\nfeasibility of the approach, we utilize a CSP solver to detect inconsistencies\nin dialogues re-lexicalized by an LLM. Our findings indicate that: (i) CSP is\neffective to detect dialogue inconsistencies; and (ii) consistent dialogue\nre-lexicalization is challenging for state-of-the-art LLMs, achieving only a\n0.15 accuracy rate when compared to a CSP solver. Furthermore, through an\nablation study, we reveal that constraints derived from domain knowledge pose\nthe greatest difficulty in being respected. We argue that CSP captures core\nproperties of dialogue consistency that have been poorly considered by\napproaches based on component pipelines.","PeriodicalId":501033,"journal":{"name":"arXiv - CS - Symbolic Computation","volume":"25 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Evaluating Task-Oriented Dialogue Consistency through Constraint Satisfaction\",\"authors\":\"Tiziano Labruna, Bernardo Magnini\",\"doi\":\"arxiv-2407.11857\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Task-oriented dialogues must maintain consistency both within the dialogue\\nitself, ensuring logical coherence across turns, and with the conversational\\ndomain, accurately reflecting external knowledge. We propose to conceptualize\\ndialogue consistency as a Constraint Satisfaction Problem (CSP), wherein\\nvariables represent segments of the dialogue referencing the conversational\\ndomain, and constraints among variables reflect dialogue properties, including\\nlinguistic, conversational, and domain-based aspects. To demonstrate the\\nfeasibility of the approach, we utilize a CSP solver to detect inconsistencies\\nin dialogues re-lexicalized by an LLM. Our findings indicate that: (i) CSP is\\neffective to detect dialogue inconsistencies; and (ii) consistent dialogue\\nre-lexicalization is challenging for state-of-the-art LLMs, achieving only a\\n0.15 accuracy rate when compared to a CSP solver. Furthermore, through an\\nablation study, we reveal that constraints derived from domain knowledge pose\\nthe greatest difficulty in being respected. We argue that CSP captures core\\nproperties of dialogue consistency that have been poorly considered by\\napproaches based on component pipelines.\",\"PeriodicalId\":501033,\"journal\":{\"name\":\"arXiv - CS - Symbolic Computation\",\"volume\":\"25 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-07-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Symbolic Computation\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2407.11857\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Symbolic Computation","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2407.11857","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Evaluating Task-Oriented Dialogue Consistency through Constraint Satisfaction
Task-oriented dialogues must maintain consistency both within the dialogue
itself, ensuring logical coherence across turns, and with the conversational
domain, accurately reflecting external knowledge. We propose to conceptualize
dialogue consistency as a Constraint Satisfaction Problem (CSP), wherein
variables represent segments of the dialogue referencing the conversational
domain, and constraints among variables reflect dialogue properties, including
linguistic, conversational, and domain-based aspects. To demonstrate the
feasibility of the approach, we utilize a CSP solver to detect inconsistencies
in dialogues re-lexicalized by an LLM. Our findings indicate that: (i) CSP is
effective to detect dialogue inconsistencies; and (ii) consistent dialogue
re-lexicalization is challenging for state-of-the-art LLMs, achieving only a
0.15 accuracy rate when compared to a CSP solver. Furthermore, through an
ablation study, we reveal that constraints derived from domain knowledge pose
the greatest difficulty in being respected. We argue that CSP captures core
properties of dialogue consistency that have been poorly considered by
approaches based on component pipelines.