Incorporating discourse features into confidence scoring of intention recognition results in spoken dialogue systems

Abstract

This paper proposes a method for the confidence scoring of intention recognition results in spoken dialogue systems. To achieve tasks, a spoken dialogue system has to recognize user intentions. However, because of speech recognition errors and ambiguity in user utterances, it sometimes has difficulty recognizing them correctly. Confidence scoring allows errors to be detected in intention recognition results and has proved useful for dialogue management. Conventional methods use the features obtained from speech recognition results for single utterances for confidence scoring. However, this may be insufficient since the intention recognition result is a result of discourse processing. We propose incorporating discourse features for a more accurate confidence scoring of intention recognition results. Experimental results show that incorporating discourse features significantly improves the confidence scoring.

Temperature and magnetic field dependence of spin-ice correlations in the pyrochlore magnetTb2Ti2O7

“…27,28 Recent experiments on Tb 2 Ti 2 O 7 have shown signatures of the proposed QSI state, most notably pinch point scattering 29 indicative of Coulombic correlations in spin ice, and signs of a magnetization plateau for magnetic fields along [111] theoretically proposed by Molavian et al, 25 although there is a diversity in the interpretation of this latter evidence. [30][31][32][33][34][35] Our recent neutron scattering measurements 36 on single-crystalline Tb 2 Ti 2 O 7 at 70 mK have revealed short-range spin ice correlations at 1 2 , 1 2 , 1 2 wave vectors which can be ascribed to a shortrange antiferromagnetically ordered spin ice state with spin canting at an angle of ∼12 • from the local 111 axes. These static correlations observed in the elastic scattering channel were shown to be very sensitive to both applied magnetic field and temperature, although parametric studies of the field and temperature dependence of this scattering were not performed.…”

Conversational system for information navigation based on POMDP with user focus tracking

“…We adopt the logistic regression (LR)-based dialogue act tagging approach (Tur et al, 2006;Higashinaka et al, 2006). The probability of user intention o s given an ASR result of the user utterance h is defined as,…”

Naturalistic Dialogue Management for Noisy Speech Recognition

“…Confidence scores can also be continuous. They can be based on a normalized sum of confidence over components of the semantic representation of user utterances [43], or a probabilistic combination of acoustic, semantic and discourse features [44].…”