search

MiuLab / TC-Bot

User Simulation for Task-Completion Dialogues
MIT License
807 stars 295 forks source link

About data #15

Open qingyue2014 opened 5 years ago

qingyue2014 commented 5 years ago

In your paper, you said that ". The raw conversational data were collected via Amazon Mechanical Turk, with annotations provided by domain experts. In total, we have labeled 280 dialogues " But I can't find any label data in this code. Do you use any label data in this experiment?

netrookiecn commented 4 years ago

In your paper, you said that ". The raw conversational data were collected via Amazon Mechanical Turk, with annotations provided by domain experts. In total, we have labeled 280 dialogues " But I can't find any label data in this code. Do you use any label data in this experiment?

I think 280 labeled dialogues is used for generating user goals