英文

MIAM数据集概述

数据集摘要

多语言对话行为基准是一个用于训练、评估和分析自然语言理解系统的资源集合,专门设计用于口语。数据集包括英语、法语、德语、意大利语和西班牙语。它们涵盖了各种领域,包括自发言语、脚本场景和联合任务完成。所有数据集都包含对话行为标签。

支持的任务和排行榜

[需要更多信息]

语言

英语、法语、德语、意大利语、西班牙语。

数据集结构

数据实例

Dihana语料库

对于dihana配置,数据集中的一个示例是:

{
  'Speaker': 'U',
  'Utterance': 'Hola , quería obtener el horario para ir a Valencia',
  'Dialogue_Act': 9,  # 'Pregunta' ('Request')
  'Dialogue_ID': '0',
  'File_ID': 'B209_BA5c3',
}
iLISTEN语料库

对于ilisten配置,数据集中的一个示例是:

{
  'Speaker': 'T_11_U11',
  'Utterance': 'ok, grazie per le informazioni',
  'Dialogue_Act': 6,  # 'KIND-ATTITUDE_SMALL-TALK'
  'Dialogue_ID': '0',
}
LORIA语料库

对于loria配置,数据集中的一个示例是:

{
  'Speaker': 'Samir',
  'Utterance': 'Merci de votre visite, bonne chance, et à la prochaine !',
  'Dialogue_Act': 21,  # 'quit'
  'Dialogue_ID': '5',
  'File_ID': 'Dial_20111128_113927',
}
HCRC MapTask语料库

对于maptask配置,数据集中的一个示例是:

{
  'Speaker': 'f',
  'Utterance': 'is it underneath the rope bridge or to the left',
  'Dialogue_Act': 6,  # 'query_w'
  'Dialogue_ID': '0',
  'File_ID': 'q4ec1',
}
VERBMOBIL语料库

对于vm2配置,数据集中的一个示例是:

{
  'Utterance': 'ja was sind viereinhalb Stunden Bahngerüttel gegen siebzig Minuten Turbulenzen im Flugzeug',
  'Utterance': 'Utterance',
  'Dialogue_Act': 'Dialogue_Act',  # 'INFORM'
  'Speaker': 'A',
  'Dialogue_ID': '66',
}

数据字段

对于dihana配置,不同字段如下:

  • 说话者:字符串格式的说话者标识符。
  • 话语:字符串格式的话语。
  • 对话行为:话语的对话行为标签。可以是以下之一:'Afirmacion' (0) [积极反馈],'Apertura' (1) [开场白],'Cierre' (2) [闭幕白],'Confirmacion' (3) [确认],'Espera' (4) [等待],'Indefinida' (5) [未定义],'Negacion' (6) [消极反馈],'No_entendido' (7) [请求澄清],'Nueva_consulta' (8) [新请求],'Pregunta' (9) [问题] 或 'Respuesta' (10) [回复]。
  • 对话标识符:字符串格式的对话标识符。
  • 文件标识符:字符串格式的源文件标识符。

对于ilisten配置,不同字段如下:

  • 说话者:字符串格式的说话者标识符。
  • 话语:字符串格式的话语。
  • 对话行为:话语的对话行为标签。可以是以下之一:'AGREE' (0),'ANSWER' (1),'CLOSING' (2),'ENCOURAGE-SORRY' (3),'GENERIC-ANSWER' (4),'INFO-REQUEST' (5),'KIND-ATTITUDE_SMALL-TALK' (6),'OFFER-GIVE-INFO' (7),'OPENING' (8),'PERSUASION-SUGGEST' (9),'QUESTION' (10),'REJECT' (11),'SOLICITATION-REQ_CLARIFICATION' (12),'STATEMENT' (13) 或 'TALK-ABOUT-SELF' (14)。
  • 对话标识符:字符串格式的对话标识符。

对于loria配置,不同字段如下:

  • 说话者:字符串格式的说话者标识符。
  • 话语:字符串格式的话语。
  • 对话行为:话语的对话行为标签。可以是以下之一:'ack' (0),'ask' (1),'find_mold' (2),'find_plans' (3),'first_step' (4),'greet' (5),'help' (6),'inform' (7),'inform_engine' (8),'inform_job' (9),'inform_material_space' (10),'informer_conditioner' (11),'informer_decoration' (12),'informer_elcomps' (13),'informer_end_manufacturing' (14),'kindAtt' (15),'manufacturing_reqs' (16),'next_step' (17),'no' (18),'other' (19),'quality_control' (20),'quit' (21),'reqRep' (22),'security_policies' (23),'staff_enterprise' (24),'staff_job' (25),'studies_enterprise' (26),'studies_job' (27),'todo_failure' (28),'todo_irreparable' (29) 或 'yes' (30)。
  • 对话标识符:字符串格式的对话标识符。
  • 文件标识符:字符串格式的源文件标识符。

对于maptask配置,不同字段如下:

  • 说话者:字符串格式的说话者标识符。
  • 话语:字符串格式的话语。
  • 对话行为:话语的对话行为标签。可以是以下之一:'acknowledge' (0),'align' (1),'check' (2),'clarify' (3),'explain' (4),'instruct' (5),'query_w' (6),'query_yn' (7),'ready' (8),'reply_n' (9),'reply_w' (10) 或 'reply_y' (11)。
  • 对话标识符:字符串格式的对话标识符。
  • 文件标识符:字符串格式的源文件标识符。

对于vm2配置,不同字段如下:

  • 话语:字符串格式的话语。
  • 对话行为:话语的对话行为标签。可以是以下之一:'ACCEPT' (0),'BACKCHANNEL' (1),'BYE' (2),'CLARIFY' (3),'CLOSE' (4),'COMMIT' (5),'CONFIRM' (6),'DEFER' (7),'DELIBERATE' (8),'DEVIATE_SCENARIO' (9),'EXCLUDE' (10),'EXPLAINED_REJECT' (11),'FEEDBACK' (12),'FEEDBACK_NEGATIVE' (13),'FEEDBACK_POSITIVE' (14),'GIVE_REASON' (15),'GREET' (16),'INFORM' (17),'INIT' (18),'INTRODUCE' (19),'NOT_CLASSIFIABLE' (20),'OFFER' (21),'POLITENESS_FORMULA' (22),'REJECT' (23),'REQUEST' (24),'REQUEST_CLARIFY' (25),'REQUEST_COMMENT' (26),'REQUEST_COMMIT' (27),'REQUEST_SUGGEST' (28),'SUGGEST' (29) 或 'THANK' (30)。
  • 说话者:字符串格式的说话者。
  • 对话标识符:字符串格式的对话标识符。

数据拆分

Dataset name Train Valid Test
dihana 19063 2123 2361
ilisten 1986 230 971
loria 8465 942 1047
maptask 25382 5221 5335
vm2 25060 2860 2855

数据集创建

策展原因

[需要更多信息]

源数据

初始数据收集和规范化

[需要更多信息]

谁是源语言生产者?

[需要更多信息]

注释

注释过程

[需要更多信息]

谁是标注者?

[需要更多信息]

个人和敏感信息

[需要更多信息]

使用数据的考虑事项

数据的社会影响

[需要更多信息]

偏差讨论

[需要更多信息]

其他已知限制

[需要更多信息]

其他信息

数据集策展人

匿名。

许可信息

本作品采用 Creative Commons Attribution-NonCommercial-ShareAlike 4.0 Unported License 许可。

引用信息

@inproceedings{colombo-etal-2021-code,
    title = "Code-switched inspired losses for spoken dialog representations",
    author = "Colombo, Pierre  and
      Chapuis, Emile  and
      Labeau, Matthieu  and
      Clavel, Chlo{\'e}",
    booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2021",
    address = "Online and Punta Cana, Dominican Republic",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.emnlp-main.656",
    doi = "10.18653/v1/2021.emnlp-main.656",
    pages = "8320--8337",
    abstract = "Spoken dialogue systems need to be able to handle both multiple languages and multilinguality inside a conversation (\textit{e.g} in case of code-switching). In this work, we introduce new pretraining losses tailored to learn generic multilingual spoken dialogue representations. The goal of these losses is to expose the model to code-switched language. In order to scale up training, we automatically build a pretraining corpus composed of multilingual conversations in five different languages (French, Italian, English, German and Spanish) from OpenSubtitles, a huge multilingual corpus composed of 24.3G tokens. We test the generic representations on MIAM, a new benchmark composed of five dialogue act corpora on the same aforementioned languages as well as on two novel multilingual tasks (\textit{i.e} multilingual mask utterance retrieval and multilingual inconsistency identification). Our experiments show that our new losses achieve a better performance in both monolingual and multilingual settings.",
}

贡献

感谢 @eusip @PierreColombo 添加此数据集。