AI Chat AI Image Generator AI Video Text to Speech

YARBUS : Yet Another Rule Based belief Update System

07/24/2015

∙

by Jérémy Fix, et al.

∙

∙

We introduce a new rule based system for belief tracking in dialog systems. Despite the simplicity of the rules being considered, the proposed belief tracker ranks favourably compared to the previous submissions on the second and third Dialog State Tracking challenges. The results of this simple tracker allows to reconsider the performances of previous submissions using more elaborate techniques.

Jérémy Fix
3 publications
Herve Frezza-buet
1 publication

page 12

page 13

page 14

research

∙ 10/13/2015

Hybrid Dialog State Tracker

This paper presents a hybrid dialog state tracker that combines a rule b...

0 Miroslav Vodolán, et al. ∙

research

∙ 02/21/2017

Hybrid Dialog State Tracker with ASR Features

This paper presents a hybrid dialog state tracker enhanced by trainable ...

0 Miroslav Vodolán, et al. ∙

research

∙ 09/17/2020

Zero-shot Multi-Domain Dialog State Tracking Using Descriptive Rules

In this work, we present a framework for incorporating descriptive logic...

0 Edgar Altszyler, et al. ∙

research

∙ 06/23/2015

Multi-domain Dialog State Tracking using Recurrent Neural Networks

Dialog state tracking is a key component of many modern dialog systems, ...

0 Nikola Mrkšić, et al. ∙

research

∙ 07/14/2015

Recurrent Polynomial Network for Dialogue State Tracking

Dialogue state tracking (DST) is a process to estimate the distribution ...

0 Kai Sun, et al. ∙

research

∙ 05/29/2018

Fully Statistical Neural Belief Tracking

This paper proposes an improvement to the existing data-driven Neural Be...

0 Nikola Mrkšić, et al. ∙

research

∙ 11/26/2018

Learning Latent Beliefs and Performing Epistemic Reasoning for Efficient and Meaningful Dialog Management

Many dialogue management frameworks allow the system designer to directl...

0 Aishwarya Chhabra, et al. ∙