%PDF-1.2
%
1 0 obj
<<
/Type /Catalog
/Pages 2 0 R
/Metadata 3 0 R
>>
endobj
4 0 obj
<<
/Producer (AFPL Ghostscript 6.50)
/Subject (This paper proposes a new application of automatically trained decision trees to derive the interpretation of a spoken sentence. A new strategy for building structured cohorts of candidates is also described. By evaluating predicates related to the acoustic confidence of the words expressing a concept, the linguistic and semantic consistency of candidates in the cohort and the rank of a candidate within a cohort, the decision tree automatically learn a decision strategy for rescoring or rejecting a n-best list of candidates representing a user's utterance. A relative reduction of 18.6% in the Understanding Error Rate is obtained by our rescoring strategy with no utterance rejection and a relative reduction of 43.1% of the same error rate is achieve with a rejection rate of only 8% of the utterances.)
/bibtex#2Faddress (Montreal, Canada)
/Author <43687269737469616E205261796D6F6E6420616E64204672E964657269632042E96368657420616E64204465204D6F72692C2052656E61746F20616E642047E972616C64696E652044616D6E61746920616E642059616E6E69636B20457374E87665>
/bibtex#2Fbibtexkey (Raymond.Bechet.ea_2004b)
/bibtex#2Fbooktitle (#ICASSP#)
/bibtex#2Fdoi (10.1109/ICASSP.2004.1326013)
/bibtex#2Ffile (icassp04.pdf:icassp04.pdf:PDF)
/Keywords (interpretation strategies, spoken language understanding)
/bibtex#2Fowner (Christian Raymond)
/bibtex#2Fpages (425-428)
/Title (Automatic learning of interpretation strategies for spoken dialogue systems)
/bibtex#2Fvolume (1)
/bibtex#2Fyear (2004)
/bibtex#2Fentrytype (Inproceedings)
>>
endobj
2 0 obj
<<
/Type /Pages
/Kids [5 0 R 6 0 R 7 0 R 8 0 R]
/Count 4
>>
endobj
3 0 obj
<<
/Type /Metadata
/Subtype /XML
/Length 9 0 R
>>
stream
This paper proposes a new application of automatically trained decision trees to derive the interpretation of a spoken sentence. A new strategy for building structured cohorts of candidates is also described. By evaluating predicates related to the acoustic confidence of the words expressing a concept, the linguistic and semantic consistency of candidates in the cohort and the rank of a candidate within a cohort, the decision tree automatically learn a decision strategy for rescoring or rejecting a n-best list of candidates representing a user's utterance. A relative reduction of 18.6% in the Understanding Error Rate is obtained by our rescoring strategy with no utterance rejection and a relative reduction of 43.1% of the same error rate is achieve with a rejection rate of only 8% of the utterances.
bibtex/address/Montreal, Canada
bibtex/bibtexkey/Raymond.Bechet.ea_2004b
bibtex/booktitle/#ICASSP#
bibtex/file/icassp04.pdf:icassp04.pdf:PDF
bibtex/owner/Christian Raymond
bibtex/pages/425-428
bibtex/volume/1
Christian Raymond
Fréderic Béchet
Renato De Mori
Géraldine Damnati
Yannick Estève
10.1109/ICASSP.2004.1326013
interpretation strategies
spoken language understanding
Automatic learning of interpretation strategies for spoken dialogue systems
2004
application/pdf
Inproceedings
Christian Raymond
Fréderic Béchet
Renato De Mori
Géraldine Damnati
Yannick Estève
Raymond.Bechet.ea_2004b
#ICASSP#
10.1109/ICASSP.2004.1326013
icassp04.pdf:icassp04.pdf:PDF
interpretation strategies, spoken language understanding
Christian Raymond
425-428
Automatic learning of interpretation strategies for spoken dialogue systems
1
2004
Inproceedings
endstream
endobj
5 0 obj
<<
/Type /Page
/MediaBox [0 0 595 842]
/Rotate 0
/Parent 2 0 R
/Resources 10 0 R
/Contents 11 0 R
>>
endobj
6 0 obj
<<
/Type /Page
/MediaBox [0 0 595 842]
/Rotate 0
/Parent 2 0 R
/Resources 12 0 R
/Contents 13 0 R
>>
endobj
7 0 obj
<<
/Type /Page
/MediaBox [0 0 595 842]
/Rotate 0
/Parent 2 0 R
/Resources 14 0 R
/Contents 15 0 R
>>
endobj
8 0 obj
<<
/Type /Page
/MediaBox [0 0 595 842]
/Rotate 0
/Parent 2 0 R
/Resources 16 0 R
/Contents 17 0 R
>>
endobj
9 0 obj
4195
endobj
10 0 obj
<<
/ProcSet [/PDF /ImageB /Text]
/ExtGState 18 0 R
/Font 19 0 R
>>
endobj
11 0 obj
<<
/Length 20 0 R
/Filter /FlateDecode
>>
stream
x\ےq|c^$Fp@
_P0oZVzt3nݤ/uɬFMPTeՕs>)cQ^}8a'vEŚV誅뛓pӉj?^kDL敻늵RWEYq^pgiڊ+3^皪i
.^__>jfT"+HSdk5w BJ}ʹM-Ic^^CKX탏&S7kt_2[ۂkڋ.
p)"JҵEb4,V*H0`d%n_{I*߸6!iQꗿz4_}2ūZ|zMm\Asqw\U~1*ط~XDzՍ[y ?D1nziF
Z3 RL(،iZapŀn,^U b|DIGaJ[hEf(آe1{ 0Z(җ{
]\v/ZZ^kUbLSqt0]