Skip to content

qiusihang/ticktalkturk

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

32 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TickTalkTurk

TickTalkTurk: Conversational Crowdsourcing Made Easy (PDF). Sihang Qiu, Ujwal Gadiraju, and Alessandro Bozzon. CSCW 2020 Demonstration.

@inproceedings{qiu2020ticktalkturk,
  title={TickTalkTurk: Conversational Crowdsourcing Made Easy},
  author={Qiu, Sihang and Gadiraju, Ujwal and Bozzon, Alessandro},
  booktitle={Conference Companion Publication of the 2020 on Computer Supported Cooperative Work and Social Computing},
  pages={1--5},
  year={2020}
}

logo

  • If you want to create a chatbot for online surveys/tasks, please visit this PAGE for chatbot settings.

  • The workflow of conversational microtask crowdsourcing is defined in ./js/client.js (complex task workflow can be customized here).

  • Download the code of the previous version HERE.

Explanation

  • An utterance is a message shown in a speech bubble (speech balloon) from either the chatbot or the user.

  • Currently, an utterance of the chatbot can be of 4 types, i.e. text, image, radio buttons, and checkboxes.

  • The user can give a response (i.e. an utterance) by either using text input (at the bottom of the UI) or using interactive utterances (i.e. buttons or checkboxes) sent by the chatbot.

  • A conversational turn starts with an utterance of the chatbot, and ends with an utterance of the user (as a response to chatbot).

  • The utterance of the user can be validated using some keywords or phrases. If a keyword/phrase provided is found in the user's utterance, the conversation proceeds to the next turn. Otherwise, the user will be asked to provide a valid response.

  • Each turn has at least one chatbot's utterance and only one user's utterance.

  • A conversation consists of turn(s), and it has at least one turn.