Toshiba AI Technology Catalog

  • Speech dialogue

Statistical dialog system

Responds appropriately to inquiries from humans.


  • Based on deep learning, the system accurately understands the spoken content.
  • Builds dialog scenarios to select the most appropriate response based on reinforcement learning.
  • Joint research with the University of Cambridge and the Toshiba Europe Cambridge Research Laboratory

Applications



  • Digital signage at shopping centers and other facilities
  • Work support at plants and other workplaces by speech
    • Work detail reports
    • Instructions for inspections
    • Searches for parts to be replaced, etc.

Benchmarks, strengths, and track record



  • The accuracy of unknown keyword detection in user speech enables the world’s highest level of performance in the evaluation using public data sets.

Inquiries



Please include the title “Toshiba AI Technology Catalog: Statistical dialog system” or the URL in the inquiry text.
Please note that because this technology is currently the subject of R&D activities, immediate responses to inquiries may not be possible.

References:

  • Y. Kobayashi, et al., “Slot Filling with Weighted Multi-Encoders for Out-of-Domain Values”, INTERSPEECH 2019.
  • K. Iwata, et al., “Transfer Learning for Unseen Slots in End-to-end Dialogue State Tracking”, IWSDS 2019.
  • T. Yoshida, et al., “Dialog State Tracking for Unseen Values using an Extended Attention Mechanism”, IWSDS 2018.