Skip to main navigation Skip to search Skip to main content

A fast training algorithm of multiple-timescale recurrent neural network for agent motion generation

  • Kyungpook National University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Motion understanding and regeneration are two basic aspects of human-agent interaction. One important function of agents is to represent human's activities. For better interaction with human, robot agents should not only do something following human's order, but also be able to understand or even play some actions. Multiple Timescale Recurrent Neural Networks (MTRNN) is believed to be an efficient tool for robots action generation. In our previous work, we extended the concept of MTRNN and developed Supervised MTRNN for motion recognition. In this paper, we use Conditional Restricted Boltzmann Machine (CRBM) to initialize Supervised MTRNN and accelerate the training speed of Supervised MTRNN. Experiment results show that our method can greatly increase the training speed without losing much performance.

Original languageEnglish
Title of host publicationHAI 2015 - Proceedings of the 3rd International Conference on Human-Agent Interaction
PublisherAssociation for Computing Machinery, Inc
Pages243-246
Number of pages4
ISBN (Electronic)9781450335270
DOIs
StatePublished - 21 Oct 2015
Event3rd International Conference on Human-Agent Interaction, HAI 2015 - Daegu, Korea, Republic of
Duration: 21 Oct 201524 Oct 2015

Publication series

NameHAI 2015 - Proceedings of the 3rd International Conference on Human-Agent Interaction

Conference

Conference3rd International Conference on Human-Agent Interaction, HAI 2015
Country/TerritoryKorea, Republic of
CityDaegu
Period21/10/1524/10/15

Keywords

  • Action generation
  • Machine learning
  • Recurrent Neural Network
  • Restricted Boltzmann Machine

Fingerprint

Dive into the research topics of 'A fast training algorithm of multiple-timescale recurrent neural network for agent motion generation'. Together they form a unique fingerprint.

Cite this