Announcing Two New Natural Language Dialog Datasets

By Google AI Blog - 2020-03-26

Description

Posted by Bill Byrne and Filip Radlinski, Research Scientists, Google Research Today’s digital assistants are expected to complete tasks...

Summary

  • Posted by Bill Byrne and Filip Radlinski, Research Scientists, Google Research Today’s digital assistants are expected to complete tasks and return personalized results across many subjects, such as movie listings, restaurant reservations and travel plans.
  • However, despite tremendous progress in recent years, they have not yet reached human-level understanding.
  • These 2-person dialogs naturally include disfluencies and errors that happen spontaneously between the two parties that are difficult to replicate using synthesized dialog.
  • Task-Oriented Dialog TheTaskmaster-1 dataset makes use of both the methodology described above as well as a one-person, written technique to increase the corpus size and speaker diversity—about 7.7k written “self-dialog” entries and ~5.5k 2-person, spoken dialogs.

 

Topics

  1. UX (0.29)
  2. NLP (0.21)
  3. Backend (0.11)

Similar Articles

A swiss cheese model for reducing biases in user research

By Medium - 2020-12-27

Every accident has underlying factors behind its occurrence. Those factors are usually human or design-related which causes slips and mistakes, which cause the accident. Sometimes we detect the right…

huggingface/datasets

By GitHub - 2021-01-05

🤗 The largest hub of ready-to-use NLP datasets for ML models with fast, easy-to-use and efficient data manipulation tools - huggingface/datasets