- - PowerPoint PPT Presentation

!"#"$%&'("$)*&+$&%,"&-"."/012"$%&03& 41""#,&'"#,$0/05+"*&6$)&-+6/05&47*%"2*&

!"#$%&'($)*&+"((,-& .//*$,%&.(0$1$2$"*&340,**$5,42,&'()6/& 74$#,(-$%"%&8"(*)-&333&%,&9"%($%& & %5($)*:$41;62<=;,-& &

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, 9" :1*1$%,;&1$(7<,=(3/%3%+'$,

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

89%/+$"& :; <$%(0)9#%+0$&

• "

./'01$,2+34'5,.67%187, 9" :1*1$%,;&1$(7<,=(3/%3%+'$,

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

#$%&'()*%+'$,

>,?&%,#$2,-&0)&"22,--&$41)(="0$)4&$4&0@,&%$5$0"*&%)="$4& & & >,2,--"(A&0)&/()#$%,&-0("$5@01)(?"(%&?"A-&)1&(,0($,#$45&$0;&& & & BCD,20$#,E&,11,20$#,F&,"-AF&-"1,&"4%&0("4-/"(,40&$40,("20$)4& C,0?,,4&0@,&6-,(&"4%&0@,&-A-0,=;&& & & G$=6*"0$)4&)1&&-/)H,4&@6="4I0)I@6="4&2)==64$2"0$)4;& !"#$%&'()*+,-*&

89%/+$"& !" #$%&'()*%+'$, =; 410>"$&-+6/05&47*%"2*& 9" :1*1$%,;&1$(7<,=(3/%3%+'$,

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

./'01$,2+34'5,.67%187,

!"#$%&'*)*+,-E& ="4I="2@$4,& $40,(1"2,& 0@"0& $-& "C*,& 0)& (,2)54$J,& "4%&64%,(-0"4%&"&-/)H,4&$4/60&"4%&0)&/()%62,&"4&)("*&)60/60&"-&"4& "4-?,(;&& !$11,(,40&=)%6*,-&6-6"**A&0"H,&/"(0&0)&2"((A&)60&0@$-&1$4"*&5)"*;& G/)H,4&2)4#,(-"0$)4"*&"5,40-&"(,&5,4,("**A&C"-,%&)4E& 8)==64$2"0$)4&2@"44,*E&0,*,/@)4,&*$4,F&=$2()/@)4,;& K@,&0"-H&$-&(,-0($20,%&0)&"&2,(0"$4&H4)?*,%5,&%)="$4;& 9$L,%&$4$0$"0$#,&-0("0,5A;& >"06("*&*"456"5,&$4/60-&"(,&"22,/0,%M5,4,("0,%;& .//$"0#+"%1*2'K@,&2)=/*,L$0A&)1&0@,&2)4#,(-"0$)4&2"4&#"(A&1()=&"&

• $=/*,&)/0$)4&2@))-$45&$4&"&=,46F&0)&"&2)4#,(-"0$)4"*&2)=/"4$)4;&&

&

./'01$,2+34'5,.67%187,

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

?11/+#6%+0$*&%,(095,&"@621/"*&&

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

./'01$,2+34'5,.67%187,

?11/+#6%+0$*&%,(095,&"@621/"*&,

G)=,&,L"=/*,-E& & .%"/0&I&96*0$=)%"*&-/)H,4&%$"*)5&-A-0,=&1)(&C()?-$45& "/"(0=,40-&)4&0@,&G0)2H@)*=&(,"*&,-0"0,&="(H,0&& .=$0$,-&/()D,20&I&.60)="0,%&96*0$*$456"*&340,("20$)4&?$0@& 341)(="0$)4&"4%&G,(#$2,-&& !.NW.&8)==64$2"0)(E&340,**$5,40&2)4#,(-"0$)4"*&$40,(1"2,-&0)& %$-0($C60,%&$41)(="0$)4&& !$-8)Q&G/)H,4&!$"*)5&GA-0,=&1)(&8)41,(,42,&Q,*/&& G="(0[)=&',(="4&=6*0$=)%"*&%$"*)5&-A-0,=;&G,#,("*& "//*$2"0$)4&%)="$4-&$42*6%$45&2$4,="&C))H$45&& 8KKI+"4H&&G/,,2@&2)40()**,%&0,*,/@)4,&C"4H$45&-A-0,=;&

'

./'01$,2+34'5,.67%187,

?11/+#6%+0$*&%,(095,&"@621/"*&,

G)=,&,L"=/*,-E& G/)H,4&$40,("20$)4&2"4&C,&0@,&)4*A&?"A&0)&"22,--&$41)(="0$)4&$4&

• )=,&2"-,-F&,;5;F&0@,&-2(,,4&$-&0))&-="**&0)&%$-/*"A&$41)(="0$)4&

O,;5;&@"4%@,*%&%,#$2,-P&)(&0@,&,A,-&)1&0@,&6-,(&"(,&C6-A&$4&)0@,(& 0"-H-&O,;5;&%($#$45P& & .0@)-="$*&I&"&=6*0$*$456"*&"%"/0$#,&-/)H,4&%$"*)5&-A-0,=&1)(&,& ="$*&=,--"5,&(,"%$45&)4&"&=)C$*,&/@)4,&!38B&I&"&=6*0$=)%"*& %$"*)5&-A-0,=&0@"0&*,0-&0@,&%($#,(&2)40()*&%,#$2,-&"4%&"22,--& 340,(4,0&-,(#$2,-&6-$45&4"06("*&-/,,2@;&& 97GK&I&96*0$=)%"*F&=6*0$*$456"*&$41)(="0$)4&-,(#$2,-&1)(&

• ="**&=)C$*,&K,(=$4"*-&&&

3GTIZB!3G&"4%&ZB!3GE&Z)$2,&B/,("0,%&!($#$45&341)(="0$)4& GA-0,=&& Z38BE&Z$(06"*&340,**$5,42,&8BI%($#,(&,4"C*$45&4"06("*& $40,("20$)4&C,0?,,4&@6="4-&"4%&%$5$0"*&%,#$2,-&"4%&-,(#$2,-&$4& 0@,&2"(;&&

'

./'01$,2+34'5,.67%187,

?11/+#6%+0$*&%,(095,&"@621/"*&,

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

./'01$,2+34'5,.67%187,

?11/+#6%+0$*&%,(095,&"@621/"*&, SL"=/*,&)1&=6*0$=)%"*&%$"*)5&-A-0,=-&

G="(0H)=& & !"#$%&"'(%'()"'*&+,(-%&'./0%,&+(.%/'121("&3'4'+&' *&+,(+5613'7%8'&+2'4')"#9'2%6:' !)+('.1'9#+2./;'+('()"'$./"&+'(%/.;)(:' <)"1"'+,"'()"'&%=."1'9#+2./;'(%/.;)(3'<)"'$./"&+1')+="' >""/'&+,5"?'%/'()"'&+9'

'

./'01$,2+34'5,.67%187,

, ./'01$,2+34'5,.67%187,

34&0@,&\Z33&2,406(AF&-)=,&/@$*)-)/@,(-&*$H,& !,-2"(0,-&)(&T,$C4$J&C,5"4&?$0@&0@,&2)4-0(620$)4&)1& 64$#,(-"*&*"456"5,-&$4&1)(=&)1&46=,($2"*&2)%,-;&& & & & & V$(-0&-,($)6-&"00,=/0-&0)&C6$*%&0"*H$45&-A-0,=-& ?@,(,&$4$0$"0,%&$4&0@,&\Z333&"4%&\3\&2,406($,-;& V$(-0&"60)="0"&"C*,&0)&$=$0"0,&@6="4&C,@"#$)6(& ?,(,&2)4-0(620,%;&

Hello

• Hola
• 1

2 3

• ./'01$,2+34'5,.67%187,

A(+"3&,+*%0(+#6/&(".+"B&&

,

+"()4&Z)4&[,=/,*,4&& ]^^_&

K"*H$45&@,"%-& V$(-0&"60)="0"&0)&/()%62,& ?@)*,&?)(%-&"4%&-@)(0& /@("-,-;& &

Y)-,1&V"C,(&&]`a^&

3=$0"0$)4&)1&@6="4&-/,,2@& /()%620$)4& S6/@)4$";&]b&H,A-&0)&1)(=& "4A&?)(%& &

./'01$,2+34'5,.67%187,

A(+"3&,+*%0(+#6/&(".+"B&&

& .0&0@,&,4%&)1&\3\&2,406(AF&-2$,40$-0-&O,;5;&Z)4& Q,(*=@)*0JP&2)42*6%,%&0@"0&-/,,2@&2)6*%&C,& /()%62,%&,*,20($2"**A&$4-0,"%&)1&=,2@"4$2"**A;& & & Yc&G0,?"(0&$4&0@,&C,5$44$45&)1&\\&2,406(A&C6$*0& "&="2@$4,&0@"0&2)6*%&5,4,("0,&#)2"*$2&-)64%-& ,*,20($2"**A;&&

This was the theory, what about the practice??

./'01$,2+34'5,.67%187,

56*'

V$(-0&,*,20($2&="2@$4,-F&1)(&,L"=/*,E&

ZB8B!SNF&-/,,2@&"4"*AJ,(&"4%&-A40@,-$J,(&%,#,*)/,%&$4&+,**& T"C)("0)($,-&0@"0&2)6*%&C,&)/,("0,%&CA&"&H,AC)"(%;& V$(-0&0("4-*"0$45&="2@$4,-&"(,&/"0,40,%&&V("42,&O"60)="0$2F&

• $=/*,P&"4%&N6--$"&O"%#"42,%F&@6="4&$40,(#,40$)4P;&

76*'

V$(-0&2)=/60,(-&?,(,&%,#,*)/,%&%6($45&XXd&

]e<`&&]ef]E&[)(4"%&g6-,F&1$(-0&/()5("=="C*,&2)=/60,(;& ]ef<E&8)*)--6-F&+*,02@*,A&W"(H;& ]ef<&&]efaE&74$#;&W,44-A*#"4$"F&S>3.8;&

K@,&/)0,40$"*&)1&0@,-,&="2@$4,-&C,5$4-&0)&C,& 2)4-$%,(,%&1)(&)0@,(&/6(/)-,-E&

.*"4&K6($45& X"((,4&X,"#,(&

&

&

./'01$,2+34'5,.67%187,

86*'

a_&I&.*"4&K6($45&&340,**$5,40&9"2@$4,-& a]&&8))/,(F&T$C,(="4&"4%&+)(-0&&G/,20)5("=& ad&&9"2@$4,&0("4-*"0$)4&2)41,(,42,-& a^&&8@)=-HA&&GA40"20$2&G0(6206(,-& a`&&W,0,(-)4F&X"45&"4%&G$#,(-0,4&&!$/@)4,-&

96*'

V$(-0&2)4#,(-"0$)4"*&"5,40-F&,;5;&X,$J,4C"6=&ST3g.& &>)0&2)=/60"0$)4"*&*$456$-0$2-&

:6*'

8)=/60"0$)4"*&*$456$-0$2& +"-,%&$4&0@,)($2"*&?)(H&"C)60&-,="40$2&"4%&*"456"5,& 2"(($,%&)60&CA&8@)=-HAF&9)40"56,F&X))%& N6*,&C"-,%&-/,,2@&-A40@,-$-&

&

./'01$,2+34'5,.67%187,

;6*'

K,*,/@)4A&-,(#$2,-& !$20"0$)4&-A-0,=-&

<6*'

8)(/)("&%,#,*)/=,40E&X)(%4,0& 3-)*"0,%&?)(%&(,2)5;&K,*,/@)4A&%$"*)5&-A-0,=-& !,%$2"0,%&2)=/"4$,-;&

=666*'

G0"0$-0$2"*&-A-0,=-&&N6*,-&-0$**&$4&/("20$2"*&6-,& G0"4%"(%-&1)(&,"-A&%,#,*)/=,40;&

=6>6*'

7-"C$*$0AF&6-,(&2,40(,%&-A-0,=-;& &

&

./'01$,2+34'5,.67%187,

./'01$,2+34'5,.67%187,

?(#,+%"#%9("&

ASR NLP DM NLG SS User voice System voice

INPUT INT ERFA C E OUTPUT INT ERFA C E

./'01$,2+34'5,.67%187,

?(#,+%"#%9("&

Conversational Agent Integration of the different agents

./'01$,2+34'5,.67%187,

?(#,+%"#%9("&

?9%026%+#&41""#,&!"#05$+%+0$& G)=,&.GN&2)=/*,L$0$,-E&

T$456$-0$2&#"($"C$*$0A& 340,(I-/,"H,(&#"($"C$*$0A& 340("I-/,"H,(&#"($"C$*$0A& 8@"44,*&#"($"C$*$0A&

& & 9"AC,&A)6&2)6*%&"%%&A)6(&)?4&0)&0@,&*$-0RR& ASR NLP User voice

./'01$,2+34'5,.67%187,

(+#+"*+"0#$'.//?%#0@2''A#?B%C'D"33,1'A%3,$*' BCD,20$#,E&K)&1$4%&0@,&-,h6,42,&)1&?)(%-&X&5$#,4&"&

• ,h6,42,&)1&"2)6-0$2&%"0"&.E&

& & & & & K("$4$45&%"0"i& & & & &

./'01$,2+34'5,.67%187,

Acoustic Model Language Model

?9%026%+#&41""#,&!"#05$+%+0$&

C6%9(6/&D6$5965"&E(0#"**+$5&

ASR NLP DM Levels of NLP:

• Morphological
• Lexical
• Sintactical
• Semantic
• Discourse
• Pragmatical

It has many applications, among them: Machine translation Question answering, text summarization, text simplification Information classification, filtering and recovering, e.g. spam filters, orthographic correctors Dialog systems In dialog systems: Obtains the meaning of the recognized user utterance In a wider sense: Understanding human natural language

./'01$,2+34'5,.67%187,

K@,&5)"*&)1&-,="40$2&"4"*A-$-&$-&0)&(,/(,-,40&0@,&=,"4$45&)1& 0@,&-,40,42,-&$4&"&*)5$2"*&*"456"5,&

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

IJK,E,2+7*')&71,

342(,=,40"*&340,(/(,0"0$)4& K@,&%$-2)6(-,&$-&"&-,h6,42,&)1&-,40,42,-&0@"0&2"4&C,&-0(6206(,%&$4&"& ?"A&)(&"4)0@,(&0)&/()%62,&%$11,(,40&=,"4$45-;& N@,0)($2& N,1,(,42,&(,-)*60$)4& .4"/@)("& QA/)4A=-& W()4)64-& >>-&O/;,;&X@"0&?"-&/(,#$)6-*A&2)==,40,%P;& B0@,(&,L/(,--$)4-&O/;,;&@,&?"-&1,,%$45&0@,&/$5-P;& X)(*%&H4)?*,%5,&4,,%,%& G6//)-$0$)4-&O,;5;&3&2"=,&$4&=A&C$H,;&K@,&?@,,*&="H,-&"&-0("45,& 4)$-,P;& &

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

IJK,E,=8?+5)+%6,

E%-/$,F"+)'%G'1#+4?#$'$#1&4#&,'"*'"1'"+*'#-H"&4"+)I'J#+4?#$' $#1&4#&,'"*'#-H"&4%4*'#+'C#?"%4*'$,C,$*E& & T,L$2"*&& G"=,&?)(%&=$5@0&@"#,&%$11,(,40&5("=="0$2"*&2"0,5)($,-;& GA40"20$2& W@("-,-&?$0@&=)(,&0@"4&)4,&"--)2$"0,%&-A40"20$2&-0(6206(,;& G,="40$2& G"=,&?)(%&%$11,(,40&=,"4$45-;& G"=,&-A40"20$2&-0(6206(,F&%$11,(,40&=,"4$45-;& !$-2)6(-,& N,1,(,40$"*& W("5="0$2& &

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

B("*&4"06("*&*"456"5,&@"-&/,26*$"($0$,-&0@"0& ="H,&$0&,#,4&=)(,&%$11$26*0&0)&/()2,--E&

TS\38.T&

G,#,("*&6-,(-&0"*H$45&"0&-"=,&0$=,&&G,/"("0$)4& Q)=)/@)4$2&?)(%-& 8)**)h6$"*&1)(=&)1&?)(%-& V("5=,40-F&@,-$0"0$)4-F&(,/,0$0$)4-& c6"-$*,L$2"*&O,;5;&"@"P&

9BNWQBTB'38.T&

Q,-$0"0$)4-& X)(%&2*"--,-&0@"0&)4*A&"//,"(&$4&)("*&*"456"5,F&*$H,&1)(& ,L"=/*,&%$-2)6(-,&="(H,(-&O,;5;&?,**P& 340,(D,20$)4-& 96*0$I?)(%&,L/(,--$)4-&O,;5;&3&-,,P&

&

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

GA40"20$2&

342)=/*,0,&-0(6206(,-&

@,"'2%6'%5:' ' !%,,."?:' A"1'

N,/"("0$)4-&

B/'C%/?+2D'%/'<6"1?+2D'4'8"/('(%'()"''>"+$)'

N,/,0$0$)4-&

C2'1.1(",D'&2'1.1(",'.1'=",2'0,."/?#2'

GA40"20$2&2)=C$4"0$)4-&

'

'

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

K%**"H$,'"-/$,-,1+#+"%1*' L4$,MH#*,3'-,+@%3%$%&",*I' SL0("20&0@,&-,="40$2&$41)(="0$)4&1()=&"&-A40"LI-,="40$2& "4"*A-$-&)1&0@,&-,40,42,-E&

G/,2$1$2&5("=="(-&%,1$4,%&1)(&0@,&0"-H;&& !,0,20$)4&)1&-$45*,M-,h6,42,-&)1&H,A&?)(%-&?$0@&

• ,="40$2&=,"4$45;&&

(+#+"*+"0#$'-,+@%3%$%&",*I' !,1$4$0$)4&)1&*$456$-0$2&64$0-&?$0@&-,="40$2&2)40,40&"4%& 2)4-0(620$)4&)1&"&-0"0$-0$2"*&=)%,*&1()=&"&2)(/6-&)1&*"C,*,%&

• "=/*,-;&

K("$4$45&0@,&=)%,*E&&8"/06(,&2)((,-/)4%,42,-&C,0?,,4&$4/60& 0,L0&"4%&$0-&2)((,-/)4%$45&-,="40$2&$40,(/(,0"0$)4;& K("$4$45&%"0"ii&

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

B2BCLI,/&'M1*%,,

• USERS: Semantic interpretation using frames (concepts and attributes).
• Task-dependent concepts: Availability, Booking, Booked and Cancellation.
• Task-independent concepts: Acceptance, Rejection and Not-Understood.
• Attribute: Sport, Hour, Date, Court-Type, Court-Number and Order-

Number. Semantics of the task: Yes, I want to book a tennis court for tomorrow afternoon.

(Affirmation) (Booking) Sport: tennis Date: tomorrow Hour:afternoon

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

• SYSTEM: Semantic representation using frames.
• Concepts: Availability, Booking, Booked, Cancellation, Sport, Date, Hour,

Court-Type, Confirmation-Availability, Confirmation-Booking, Confirmation- Booked, Confirmation-Cancellation, Confirmation-Sport, Confirmation-Date, Confirmation-Hour, Confirmation-CourtType, Rule-Info, Booking-Choice.

• Atributtes: Sport, Hour, Date, Court-Type, Court-Number, Order-Number y

Availability-Number. Semantics of the task: Do you want to book the squash court number 1 for February 2nd?

(Confirmation-Booking) Sport: squash Date: 02-02-2012 Court-Number:1

./'01$,2+34'5,.67%187,

B2BCLI,/&'M1*%,,

C6%9(6/&D6$5965"&E(0#"**+$5&

1 2

slots frames

3 4 5 6 7 8 n

Process of unification of frames: to combine several frames to generate a new frame . . . . . .

./'01$,2+34'5,.67%187,

C6%9(6/&D6$5965"&E(0#"**+$5&

<a& DEST

2 3

6 ORG S Tell me the origin city

1

U: I want to go to Madrid

Madrid

HOUR S: There are four buses from Valencia to Madrid at 9, 10, 11and12. Tell me the departure hour U: At 12 Unification frames 1 and 2 # FR 7 Unification frames 5 and 6 S: Do you want a direct bus? TYPE U: Yes

Direct

4

12:002 Valencia Madrid Direct 12

5 Unification frames 3and 4

Valencia Madrid Direct

STAT S: Welcome to the system... U: From Valencia

Valencia Valencia Madrid

(Viajero system)

35

./'01$,2+34'5,.67%187,

12:002

NLP DM NLG

• Deals with different sources of information
• NLP results, database query results
• Application domain World knowledge
• Knowledge about users and their intentions
• Dialog history
• Depends on:
• Task
• Flexibility
• Initiative

Inteligence Decision - Coordination

complex

• +6/05&26$65"2"$%&

./'01$,2+34'5,.67%187,

K@,(,&$-&4)0&"4&"5(,,=,40&"C)60&?@"0&0@,&0"-H&)1&"& %$"*)5&="4"5,(&"(,;&Q,(,&0@,(,&$-&"&-@)(0&*$-0E&

!,2$%,&0@,&4,L0&-A-0,=&06(4&"4%&?@$2@&$41)(="0$)4&0)&/()#$%,;&

X@"0&$41)(="0$)4&0)&/()#$%,R&Q)?R&& X@,(,&0)&5,0&0@,&$41)(="0$)4R&9"4"5,&=6*0$/*,&-)6(2,-&& Q)?&$-&0@,&%,2$-$)4&0"H,4R&V()=&*"-0&6-,(&06(4R&V()=&?@)*,&%$"*)5& @$-0)(AR&

9"4"5,&$40,("20$)4&

X@)&@"-&0@,&$4$0$"0$#,&$4&0@,&%$"*)5R&&

9"4"5,&,(()(-&$4&(,2)54$0$)4&)(&64%,(-0"4%$45&=)%6*,-&

3-&0@,&$4/60&0)&0@,&="4"5,(&2)((,20R& 31&4)0F&@)?&0)&2)((,20&$0R&

9"4"5,&2)41$(="0$)4-&

X@$2@&/$,2,-&)1&$41)(="0$)4&"(,&4,2,--"(A&0)&2)41$(=R& Q)?R&&

[,,/&0("2H&)1&-A-0,=&@$-0)(A&

X@$2@&$41)(="0$)4&$-&4,2,--"(A&0)&-0)(,R& Q)?R& &

./'01$,2+34'5,.67%187,

• +6/05&26$65"2"$%&

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

./'01$,2+34'5,.67%187,

• +6/05&26$65"2"$%&

34$0$"0$#,&2"4&C,E&

GA-0,=&%$(,20,%& 7-,(&%$(,20,%&O/$)4,,(E&Q)?&="A&3&@,*/&A)6R&.KjKP& 9$L,%& & G0"0$2&&.*?"A-&0@,&-"=,&$4$0$"0$#,&0A/,& !A4"=$2&&34$0$"0$#,&2@"45,-&%6($45&0@,&$40,("20$)4& &

8)41$(="0$)4&-0("0,5$,-E&

3=/*$2$0& SL/*$2$0& &

Control Flexibility

./'01$,2+34'5,.67%187,

• +6/05&26$65"2"$%&

C6%9(6/&/6$5965"&5"$"(6%+0$&

DM NLG SS 5 basic steps:

Content organization Discourse planning Content distribution in sentences Lexicalization Referential expressions generation Linguistic realization

Generates natural language from a machine representation of the content to be conveyed

Id-message: 000 Relation: SHOWING Arguments: Film: VÉRTIGO Room: 2 Session: 18:30

./'01$,2+34'5,.67%187,

!$-2)6(-,&/*"44$45E& G0(6206(,&"4%&)(5"4$J,&=,--"5,-&0)&C,&5,4,("0,%;& 30&2"4&@"#,&-,#,("*&*,#,*-&)1&2)=/*,L$0AE& B(5"4$J"0$)4&C"-,%&)4&/*"44$45&&.(0$1$2$"*& $40,**$5,42,F&/(,2)4%$0$)4-&"4%&,11,20-;& B(5"4$J"0$)4&C"-,%&)4&0,=/*"0,-&&G="**&46=C,(&)1& 2@)$2,-&0@"0&2"4&C,&5,4,("0,%&$4-0"40$"0$45& 0,=/*"0,-;& ' ' ' 8)4D6420$)4& S**$/-$-& +6$*%$45&)1&?)(%&-,0-& 8)=/*,L&-,40,42,-&

./'01$,2+34'5,.67%187,

C6%9(6/&/6$5965"&5"$"(6%+0$&

• K("4-1)(=-&0@,&$4/60&0,L0&$40)&

"&-A=C)*$2&(,/(,-,40"0$)4;& ]; K("4-1)(=-&46=C,(-F& "CC(,#$"0$)4-&"4%&-A=C)*-&$40)& ?)(%-;&& d; .--$54-&/@)4,0$2& 0("4-2($/0$)4-&0)&,"2@&?)(%& "4%&%$-0$456$-@,-&/()-)%$2& 64$0-&*$H,&/@("-,-&"4%&

• ,40,42,-;&&
• '("/@,=,I0)I/@)4,=,&&

'KW&

• K,L0I0)I/@)4,=,&&I&KKW&

Text pre- processing or normalization X-to-phoneme

Text with numbers, symbols, etc. Text composed

• nly by words

Symbolic linguistic representation

./'01$,2+34'5,.67%187,

C6%9(6/&/6$5965"&5"$"(6%+0$&

!$11,(,40&0,2@4$h6,-&)1&G/,,2@&GA40@,-$-E' +"-,%&$4&@6="4&-/,,2@&/()%620$)4& W"("=,0($2&&9)%,*&-,#,("*&/@A-$)*)5$2"*& /"("=,0,(-&"C)60&#)2"*&0("20&0)&/()%62,&-)64%-;& +"-,%&$4&1)(="40-&&V)(="40-&=)%,*&#$C("0$)4&)1& #)2"*&2@)(%-;&& X$0@)60&/@A-$)*)5$2"*&C"-,& 8)42"0,4"0$#,&-A40@,-$-&&Y)$4&/(,I(,2)(%,%&64$0-& )1&@6="4&#)$2,&&74$0&-$J,RR&

Naturalness Automatization

./'01$,2+34'5,.67%187,

C6%9(6/&/6$5965"&5"$"(6%+0$&

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, F; !"#"$%&'("$)*G&?)61%6%+0$&

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

&

!"#"$%&'("$)*&

W()"20$#$0A& .=3&,4#$()4=,40-& .60)="0$2&(,-)*60$)4&)1&/()C*,=-& B/0$="*&%$"*)5&-0("0,5$,-& 8)40,L0I"?"(,&-A-0,=-& S=)0$)4"*&$40,**$5,42,& W)(0"C$*$0A& 96*0$=)%"*F& =6*0$*"456"5,F=6*0$%)="$4&

• A-0,=-&

N,2)==,4%$45&-A-0,=-& 8)=/"("0$#,&,#"*6"0$)4& K$=,&"4%&,11)(0&=$4$=$J"0$)4& &

:1*1$%,;&1$(7,

ADAPTATION

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, F; !"#"$%&'("$)*G&?)61%6%+0$&

6H 4%6%+*%+#6/&)+6/05&26$65"2"$%& ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

!,-$54&)1&%$"*)5&-A-0,=E&& G/,2$1$2"0$)4&)1&&0@,&-A-0,=&"20$)4-&$4&(,-/)4-,&0)&6-,(&600,("42,-& "4%&,4#$()4=,40"*&-0"0,-&& !$"*)5&-0("0,5A&0@"0&%,1$4,-&0@,&2)4#,(-"0$)4"*&C,@"#$)(&)1&0@,&

• A-0,=;&&

& & '(,"0&,11)(0&$-&,=/*)A,%&0)&,=/$($2"**A&%,-$54&%$"*)5&-0("0,5$,-&1)(& 2)==,(2$"*&-A-0,=-&O4)0&"&0($#$"*&0"-HP&

:1*1$%,;&1$(7,

• +6/05&I6$65"2"$%&

O%"0,PAQ&OOPAQR' X<8U-&-0"4%"(%&\9T&1)(="0&1)(&-/,2$1A$45&$40,("20$#,&#)$2,& %$"*)5-&C,0?,,4&"&@6="4&"4%&"&2)=/60,(;&& 30&"**)?-&#)$2,&"//*$2"0$)4-&0)&C,&%,#,*)/,%&"4%&%,/*)A,%&$4&"4& "4"*)5)6-&?"A&0)&QK9T&1)(&#$-6"*&"//*$2"0$)4-;&& Y6-0&"-&QK9T&%)26=,40-&"(,&$40,(/(,0,%&CA&"&#$-6"*&?,C& C()?-,(F&Z)$2,\9T&%)26=,40-&"(,&$40,(/(,0,%&CA&"&#)$2,& C()?-,(;& S#)*60$)4E&

C>I6'S=666R' C=I6'S=667R' C=I>'S=66:R' O5I6'S=6>6R'

:1*1$%,;&1$(7,

http://www.w3.org/TR/voicexml20/ http://www.w3.org/TR/voicexml21/

• +6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

Voice XML 2.0 Grammar (SRGS) Synthesis (SSML) Call Control (CCXML) Semantic Interpret- Ration (SISR)

Recommendation Proposed Recommendation Candidate Recommendation Last Call Working Draft Requirements Working Draft

Voice XML 2.1 Voice XML 3.0

:1*1$%,;&1$(7,

E'S0%-/4+,?R2'T,$0%-,'+%'+@,'U,#+@,?' "1G%?-#+"%1'*,?C"0,I'T@#+'*+#+,V' D'S@4-#1R2'D,$/' E2'K$,#*,'*/,#B'+@,'*+#+,'G%?'U@"0@')%4'U#1+' +@,'U,#+@,?I' D2'W,%?&"#' E2'T@#+'0"+)V' D2'XH$"*"' E2'N'3"3'1%+'413,?*+#13'U@#+')%4'*#"3I'T@#+' 0"+)V' D2'A#0%1' E2'X@,'0%13"+"%1*'"1'A#0%1'W,%?&"#'#?,' *411)'#13'0$,#?'#+'>>'.A'III' '

• +6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

A P P L I C A T I O N

LEAF

leaf.vxml

ROOT

app- root.vxml

:1*1$%,;&1$(7,

8E&X,*2)=,&@)=,;&G"A&)4,&)1E&-/)(0-k&?,"0@,(k&G0"(5"J,(&"-0()/@A-$2-&4,?-;& QE&.-0()*)5A;& 8E&3&%$%&4)0&64%,(-0"4%&?@"0&A)6&-"$%;&O"&/*"01)(=I-/,2$1$2&%,1"6*0&=,--"5,;P&& 8E&X,*2)=,&@)=,;&G"A&)4,&)1E&-/)(0-k&?,"0@,(k&G0"(5"J,(&"-0()/@A-$2-&4,?-;& QE&-/)(0-;& 8E&O/()2,,%-&0)&@00/EMM???;-/)(0-;,L"=/*,;2)=M#L=*M-0"(0;#L=*P&& &

• +6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

GRAMMAR

SRGS - XML

• +6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

GRAMMAR

SRGS - ABNF

• +6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

Z)$2,\9T&"(2@$0,206(,&1)(&0@,&%,-$54&)1&%$"*)5&-A-0,=-&

• +6/05&I6$65"2"$%&

&

ZB\SB&8B997>3Kl&O@00/EMM2)==64$0A;#)L,);2)=P&

&

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

'

'

:1*1$%,;&1$(7,

• +6/05&I6$65"2"$%&

B42,&0@,&%$"*)5&-0("0,5A&@"-&C,,4&%,-$54,%F&0@,&$=/*,=,40"0$)4&)1& 0@,&-A-0,=&2"4&C,&2"(($,%&)60&6-$45&0@,&-0"4%"(%&Z)$2,\9T @)?&0)&%,-$54&0@$-&-0("0,5AR&& .//*$2"0$)4&)1&-0"0$-0$2"*&"//()"2@,-&0)&%$"*)5&="4"5,=,40& & & B7N&GBT7K3B>E&&8)=C$4,&0@,&1*,L$C$*$0A&)1&-0"0$-0$2"*&%$"*)5& ="4"5,=,40&?$0@&0@,&1"2$*$0$,-&0@"0&Z)$2,\9T&)11,(-F&& 340()%62,&-0"0$-0$2"*&=,0@)%)*)5$,-&1)(&0@,&%,#,*)/=,40&)1& 2)==,(2$"*&O"4%&4)0&-0($20*A&"2"%,=$2P&2)4#,(-"0$)4"*&"5,40-;&

:1*1$%,;&1$(7,

• +6/05&I6$65"2"$%&

K@,&6-,&)1&-0"0$-0$2"*&=)%,*-&0@"0&"(,&"60)="0$2"**A&*,"(40&1()=&%"0"& @"-&/()#$%,%&#,(A&$40,(,-0$45&(,-6*0-E& G/,,2@&(,2)54$0$)4;& T"456"5,&64%,(-0"4%$45;& !"#$%&'-#1#&,-,1+;& & K@,&/()C*,=&)1&2)#,("5,E& .**&0@,&/)--$C*,&-$06"0$)4-&0@"0&2"4&)226(&%6($45&"&%$"*)5&=6-0&C,& =)%,*,%& 6-$45& )4*A& 0@,& 0("$4$45& 2)(/6-& & E$#**"G"0#+"%1' /?%H$,-;& 8)4-$%,("C*,& ,11)(0& $-& 4,,%,%& 0)& "2h6$(,& "4%& *"C,*& "& %$"*)5& 2)(/6-&&.4+%-#+"0'3"#$%&'&,1,?#+"%1;&

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

• WB9!W&(,/(,-,40"0$)4E&

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• WB9!W&(,/(,-,40"0$)4E&

&

• +,*$,1&-/"2,E&

& &

• N,?"(%E&

& & & &

• Using Point-based Value Iteration or Q-learning, in combination

with a user simulator: & &

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• Most slot-filling dialog systems: the dialog manager takes its decisions based
• nly on the information provided by the user in the previous turns and its own

model (DIHANA task).

• Formal description:

Ai: the system answer at time i. Ui: semantic representation of the user turn at time i.

Dialog

Si: State of the dialog sequence at time i

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• Formal description:
• At time i, the objective of the dialog manager is to find the best system

answer Ai

• The selection is made by maximizing:

All the possible system answers.

• We establish a partition in the space of sequences of states:
• DRi: Dialog register at time i (concepts and attributes).

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• Formal description:
• For a sequence of states of a dialog, there is a corresponding sequence
• f DR:
• Two different sequences of states are considered to be equivalent if they

lead to the same DRi

• Great reduction in the number of different histories in the dialogs.
• A loss in the chronological information.

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• Formal description:
• The selection of the best Ai is given by:
• Each user turn:
• supplies the system with information about the task;
• provides other kinds of information, such as task-independent

information (Affirmation, Negation, and Not-Understood dialog acts).

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• Dialog Register representation:
• The DR is a sequence of 15 fields:
• We have assumed that the exact values of the attributes are not

significant to determine the next system answer:

• Five concepts: Hour, Price, Train-Type, Trip-Time, and Services.
• Ten attributes: Origin, Destination, Departure-Date, Arrival-Date, Departure-

Hour, Arrival-Hour, Class, Train-Type, Order-Number, and Services.

• DR = 15 length string of elements from {0,1,2}.
• 0: The concept is not activated, or the value of the attribute is not

given.

• 1: The concept or attribute is activated with a confidence score that

is higher than a given threshold.

• 2:The concept or attribute is activated with a confidence score that

is lower than the given threshold.

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

2#N=I=,/&'M1*%,,

MLP DRi-1, Si-1

STATISTICAL DIALOG MANAGER

System Response User Turn (frames) DR S A1i

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

• Example of a dialog:

System1: Welcome to the railway information system. How can I help you? A1: Opening DR: 00000-1000001000 User1: I want to go to Barcelona. U1=()

DR: 00000-1100001000 + Opening + U1 A2 = (Confirmation:Departure-Hour:Nil)

System2: Do you want to know the timetables? User2: Yes, for the Euromed train. U2=(Affirmation)

DR: 10000-1100001100 + Confirmation:Departure-Hour + U2

System3: Tell me the departure date. A3: (Question:Departure-Date:Nil)

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• Example of a dialog:

User3: Tomorrow U3=() DR: 10000-1120001100 + Question:Departure-Date + U3 A4 = Confirmation:Departure-Date

System4: Do you want to leave tomorrow?

User4: Yes U4=(Affirmation)

DR: 10000-1110001100 + Confirmation:Departure-Date + U4 A5 = Answer:Departure-Hour:Number-Trains,Train-Type,Departure-Hour New-Query

System5: There are several Euromed trains. The first one leaves at 08:54 and the last one at 23:45. Anything else? User5: No, thank you U5=(Negation)

DR: 10000-1110001100 + Answer-New-Query:Departure-Nil + U5 A6 = Closing

System6: Thanks for using this service. Have a good trip.

:1*1$%,;&1$(7,

• More complex dialog systems: the dialog manager selects the system turn

taking into account:

• the information provided by the user.
• the information generated by the module that controls the application.
• Application manager (AM):
• Performs the queries to the database.
• Verifies if the user query follows the regulations defined for the task.

The result of queries to the AM has to be considered in order to generate the following system turn.

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• First phase: Select the best request to be made to the AM.
• Second phase: Generate the final system turn.

We have established two phases for the selection of the next system turn in this type of tasks: We propose the use of a multilayer perceptron (MLP) to obtain the system answer.

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

• First phase (MLP 1): Select the best request to be made to the AM.
• Second phase (MLP 2): Generate the final system turn.

MLP 1

DRi-1, Si-1

STATISTICAL DIALOG MANAGER

System turn DR S User turn (frames)

APPLICATION MANAGER

MLP 2

AMi A1i A2i

:1*1$%,;&1$(7,

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

:1*1$%,;&1$(7,

Z)$2,\9T&"(2@$0,206(,&1)(&0@,&%,-$54&)1&%$"*)5&-A-0,=-&

Statistical Dialog Management&

:1*1$%,;&1$(7,

Z)$2,\9T&%)26=,40-&&0)&$=/*,=,40&0@,&-A-0,=&/()=/0-&

4%6%+*%+#6/&-+6/05&I6$65"2"$%&

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, F; !"#"$%&'("$)*G&?)61%6%+0$&

3> .%3%+7%+*34,(+34'5,83$35181$%, JH K*"(&I0)"/+$5& *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

To learn a Statistical Dialog Model Training Dialogs Dialog System Dialog Manager Statistical Dialog Model

:1*1$%,;&1$(7,

???

1) Wizard of Oz technique

Pros: Acquisition under control conditions, not system fully developed. Cons: Difficult to train the wizard(s) and to recruit people.

• DIHANA project:
• We acquired a corpus of 900 dialogs.
• We use a complete dialog system and the WOz simulates the role of the DM.
• If we want to develop a complete statistical DM:

To learn when a confirmation should be done confidence measures.

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

:1*1$%,;&1$(7,

EDECAN project:

• Two wizards of Oz:

ASR / NLU. DM.

Speech

Understanding

Simulator

TTS & Graphical Interface Dialog Management

Simulator

Dialog History Information System Application Manager Answer Generator

Answer Query Frame Codified Answer

Strategy

System Frame AM Answer

4.- Acquisition of a dialog corpus

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

1) First wizard: 2) Second wizard:

• Only supervises the output generated by an initial dialog model.

Speech

Error & Confidence Measure

Simulator

Understanding Editor

Correct frame UNDERSTANDING SIMULATOR Simulated Frame

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

2) Statistical user modeling

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

• Formal description:
• If the most probable user answer Ui is selected at each time i, the

selection is made using the following maximization: All the possible user answers.

• We establish a partition in the space of sequences of states:
• URi: User register at time i (concepts and attributes).

Two different sequences of states are considered to be equivalent if they lead to the same URi

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

• Formal description:
• The selection of the best Ui is given by:
• We propose using a MLP to make the assignation of a user turn:
• the input layer receives the current situation of the dialog.
• the output layer can be viewed as the a posteriori probability of

selecting the different user answers.

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

• User Register representation:
• The UR is a sequence of 15 fields:
• We have assumed that the exact values of the attributes are not

significant to determine the next user answer:

• Five concepts: Hour, Price, Train-Type, Trip-Time, and Services.
• Ten attributes: Origin, Destination, Departure-Date, Arrival-Date, Departure-

Hour, Arrival-Hour, Class, Train-Type, Order-Number, and Services.

• 0: The concept is not activated, or the value of the attribute is not

given.

• 1: The concept or attribute is activated.
• The error simulator perform error generation and the addition of confidence

measures.

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

• First, we evaluated the behavior of the original DM that was learned using the

training corpus (obtained by WOz).

• Then, we evaluated its evolution when the successful simulated dialogs were

incorporated to the training corpus.

• We defined four measures to evaluate the evolution of the dialog manager:
• #unseen: the number of unseen situations.
• #error: the number of answers provided by the DM that would cause the

failure of the dialog).

• %strategy: the percentage of answers provided by the DM that exactly

follow the strategy defined for the WOz to acquire the training corpus.

• %coherent: percentage of answers provided by the DM that are coherent

with the current state of the dialog.

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

#unseen

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

#error

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

%strategy and %coherent

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

3) Automatic Dialog Generation:

• Interaction of a user simulator and a dialog manager simulator.
• Initial model: Random selection of one of the possible answers defined for

the semantics of the task (user and system dialog acts).

• Learning: the probabilities of the answers selected by the dialog manager

during that dialog are incremented before beginning a new simulation.

• An error simulator module has been designed:

Performs error generation. Addition of confidence measures.

• Evaluation: A set of stop conditions is defined to automatically evaluate is

the dialog is successful or not.

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

USER SIMULATOR UNDERSTANDING AND ASR ERROR SIMULATOR DIALOG MANAGER USER SEMANTICS SYSTEM SEMANTICS

User Frames User Frames with Errors and Confidence Scores

DIALOG HISTORY AND STOP CONDITIONS DATABASE QUERY MANAGER

Databases User Turn - System Turn

NATURAL LANGUAGE GENERATION

System Frames

• Automatic Acquisition of a Dialog Corpus:

:1*1$%,;&1$(7,

K*"(&I0)"/+$5&6$)&4+29/6%+0$&

• Example of a dialog:

S1: (Opening) Welcome to the sport service. How can I help you? U1: (Booking) [0.9] Sport: Tennis [0.9] Date: 03-15-2011 [0.1] Hour: 08.00-09.00 [0.9] S2: (Confirmation-Date) Do you want to play on the 15th March? U2: (Negation) [0.9] Date: 03-14-2011 [0.9] S3: (Confirmation-Booking) Sport: Tennis Date: 2011-03-14 Hour: 08.00-09.00 {One court available} Do you want to book tennis court number 2? U3: Yes. S4: (Booking) Sport: Tennis Date: 2011-03-14 Hour: 08.00-09.00 (New-Query) Tennis court number 2 has been booked. Anything else? U4: (Negation) [0.9] S5: (Closing) Thank you for using the sport service. Goodbye. Objective: Booking Sport: tennis Date: tomorrow

:1*1$%,;&1$(7,

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, F; !"#"$%&'("$)*G&?)61%6%+0$&

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, #H L20%+0$&!"#05$+%+0$& (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

.GG,0+"C,'E%-/4+"1&E& GA-0,=&0@"0&"(,&"C*,&0)E& & .%"/0&0@,&)/,("0$)4&"22)(%$45&&0)&0@,=;& & W()#$%,&,=)0$)4"*&(,-/)4-,-E& & 3=/()#,&@6="4I2)=/60,(&$40,("20$)4;&

&

& &

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

• The emotion recognizer
• Based solely in acoustic and dialog information:
• firstly takes acoustic information into account to distinguish

between the emotions which are acoustically more different,

• secondly dialog information to disambiguate between those

that are more similar.

• Recognizing negative emotions that might discourage users from

employing the system again or even lead them to abort an

• ngoing dialog: anger, boredom and doubtfulness.
• Our emotion recognizer employs acoustic information to

distinguish anger from doubtfulness or boredom and dialog information to discriminate between doubtfulness and boredom.

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

L20%+0$&("#05$+%+0$&

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

:1*1$%,;&1$(7,

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

:1*1$%,;&1$(7,

L20%+0$&("#05$+%+0$&

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, F; !"#"$%&'("$)*G&?)61%6%+0$&

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, )H M0$%"@%N6B6("&*7*%"2& 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

@71&7,*3$,3**177,%O1,O)51,38')$%,'P,+$P'&83%+'$,3$(, F1?,71&Q+*17<,

P&'8,348'7%,1Q1&6FO1&1R, %O&')5O,(+PP1&1$%,*'88)$+*3%+'$,8'(34+%+17",, ,

BPP1*%+Q146,?&+$5+$5,+$P'&83%+'$,3$(,71&Q+*17,%',/1'/41, *3$,'$46,?1,('$1<,

7)//'&%+$5,/1&Q37+Q1,*'8/)%+$5R, *'$%1D%)34,+$P'&83%+'$,8)7%,?1,*'$Q1$+1$%46,53%O1&1(,3$(, /&'*1771(",

M0$%"@N6B6("&*7*%"2*&

:1*1$%,;&1$(7,

=8?+1$%,#$%144+51$*1,S=8#>,836,7)//'&%,%O+7,(17+&3?41, P13%)&1,'P,/1&7'$34+T3%+'$,+$,F1?,71&Q+*17<,

K1'/41,7)&&')$(1(,?6,3,71%,'P,+$%1&*'$$1*%1(,*'8/)%+$5, 3$(,71$7+$5,(1Q+*17,+$Q+7+?41,%',%O18<,

:1*'5$+T1,%O1,7)&&')$(+$5,1$Q+&'$81$%R, U3*+4+%3%1,%O1,/&'Q+7+'$,'P,/1&7'$34+T1(,71&Q+*17,%O&')5O, +$%144+51$%,3$(,+$%)+%+Q1,+$%1&P3*17",

, .1&Q+*17,/1&7'$34+T1(,3$(,(6$38+*3446,3(3/%1(,%',%O1, +$%1&3*%+'$,*'$%1D%", @71&7,$3%)&3446,/1&*1+Q1,%O1,71&Q+*1,+$%1&P3*1V,?6, 813$7,%O16,+$%1&3*%V,37,+$%144+51$%",

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

W()/)-,%&=6*0$I"5,40&"(2@$0,206(,& &

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

C'$%1D%,A3$351&!"

C&13%1V,)/(3%1,3$(,4'3(,%O1,*'$%1D%,+$P'&83%+'$,?6,813$7, 'P,7/1*+P+*,/&'P+417,377'*+3%1(,%',13*O,)71&", C'88)$+*3%1,%O+7,+$P'&83%+'$,%',%O1,2.,8'()417,()&+$5, %O1,+$%1&3*%+'$",,

C'$%1D%,K&'P+41<,

W1$1&34,)71&,+$P'&83%+'$<,

43$5)351V,70+44,41Q14,FO1$,+$%1&3*%+$5,F+%O,(+34'5,767%187V, ,

W1$1&34,7%3%+7%+*7<,$)8?1&,'P,(+34'57,3$(,(+34'5,%)&$7V,%O1+&,

()&3%+'$7V,%O1,(3%1,'P,%O1,437%,+$%1&3*%+'$,F+%O,%O1,767%18V,1%*",

@7351,7%3%+7%+*7,3$(,)71&,/&+Q+41517<,7/1*+P+*,('83+$,

0$'F41(51",

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

!" ,

• "

X'$1,$381,+(1$%+P+*3%+'$,F+%O,O+7,4'*3%+'$,3$(,3,71%,'P,*3%15'&+17, 'P,71&Q+*17,3&1,/&'Q+(1(,37,+*'$7,'$,O+7,W@#R, 9" ;O1,)71&,(1*+(17,%',)71,%O1,7/'01$,*'88)$+*3%+'$,+$%1&P3*1, G" ;O1,=CB,71$(7,3$,YAJ,/3*0351,%',%O1,2.,C'$%1D%,A3$351&<, +(1$%+P+*3%+'$,3$(,*)&&1$%,4'*3%+'$",, H" ;O1,2.,C'$%1D%,A3$351&,7141*%7,%O1,)71&,/&'P+41,3$(, *'88)$+*3%17,%O+7,+$P'&83%+'$,%',%O1,(+PP1&1$%,8'()417,'P,%O1,2.",,

E(0#"**&30//0B")&%0&6)61%&%,"&*"(.+#",

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

,

Z" B3*O,8'()41,)717,%O+7,+$P'&83%+'$,%',4'3(,+%7,7/1*+P+*, +$P'&83%+'$,3$(,8'(147,, [" ;O1,)71&,7%3&%7,%O1,+$%1&3*%+'$,F+%O,%O1,2.",, \" ;O1,#$%&'(&")*%*+',"-'%.-"&/'"0*12'"$3"&/'"$1F,P13%)&17,'$46, %',%O1,8'()417,%O3%,&1])+&17,7)*O,+$P'&83%+'$R, ^" =%,%O1,1$(,'P,%O1,+$%1&3*%+'$V,%O1,)71&,/&'P+41,+7,)/(3%1(,)7+$5, %O1,+$P'&83%+'$,'P,%O1,437%,(+34'5,7177+'$",

E(0#"**&30//0B")&%0&6)61%&%,"&*"(.+#",

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

;370<,217+5$,3$,3(3/%+Q1,767%18,%O3%,/&'Q+(17,&3+4F36, +$P'&83%+'$,+$,$3%)&34,43$5)351", @71&7,(+34'5,3*%7<,

;370E(1/1$(1$%<,4$2,5"6,78'5"9,*7%:9;<'5"9,7<:97='5"*%." >',078'-?"" 9*-@:7%.'<'%.'%&<,A337,=*&7$%5"B'+*&7$%5"B$&:C%.',-&$$.?"

A&&,7D2&'-"%''.'.!""

E,7+7%5"F'-&7%*&7$%5"F'<*,&2,':F*&'5"A,,70*1:F*&'5"978@'&: #1*--5"F'<*,&2,':4$2,5"A,,70*1:4$2,5"9,*7%:9;<'5"E,.',: B2=D',5"*%.">',078'-?""

>;-&'="&17/'$717<,,

E<'%7%+5"#1$-7%+5"B$&:C%.',-&$$.5"G*7&7%+5"B'H:I2',;5" A88'<&*%8'5"J'K'8&7$%5"I2'-&7$%5"#$%37,=*&7$%5"*%."A%-H',?"

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

4:<,_14*'81,%',%O1,&3+4F36,+$P'&83%+'$,767%18",N'F,*3$,, ,,,,,,#,O14/,6')`,,," K:<,#,F3$%,%',0$'F,%+81%3?417,%',;'&&14'('$17", 4=<,U&'8,FO+*O,7%3%+'$,(',6'),F3$%,%',413Q1`," K=<,=%'*O3,.%3%+'$", 4F<,K41371V,%144,81,(3%1",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,," KF<,;'(36", 4O<,;O1&1,3&1,P+Q1,%&3+$7,413Q+$5,P&'8,=%'*O3,.%3%+'$,%',, ,,,,,,;'&&14'('$17,3%,!a<aa38V,!!<9a38V,!H<aa/8,3$(,, ,,,,,,!^<aa/8",=$6%O+$5,1471`,,,,,,,,,,,,,,,,," KOG&;144,81,%O1,/&+*1,'P,%O1,P+&7%,'$1", 4PG&K41371V,%144,81,%O1,%+*01%,*4377",," KPG&b)7+$177,*4377", 4QG&;O1,/&+*1,+$,?)7+$177,*4377,+7,-"H,1)&'7",=$6%O+$5,1471`, KQG&I'V,%O3$0,6')", 4RG&;O3$0,6'),P'&,)7+$5,%O1,&3+4F36,+$P'&83%+'$,767%18",, ,,,,,,N3Q1,3,5''(,%&+/c, ,

B$"<,$=<&-" <',-$%*17L*&7$%?" " 9/','"7-"%$&"7%3$,=*&7$%" *D$2&"&/'"82,,'%&"1$8*&7$%" $3"&/'"2-',?" " B$".*&'"-&$,'.?" " B$"7%3$,=*&7$%"*D$2&"&/'" 2-',"<,'3','%8'-!" &7='&*D1'-?" " B$"7%3$,=*&7$%"*D$2&"&/'" 2-',"<,'3','%8'-!"&78@'&" 81*--?"

• +6/05&B+%,09%&+$#/9)+$5&#0$%"@%&+$30(26%+0$,

:1*1$%,;&1$(7,

4:<,W''(,8'&$+$5,d'7eV,#,*'$7)4%,6'),%&3+$7, +$P'&83%+'$,413Q+$5,P&'8,=%'*O3,.%3%+'$",2',6'), F3$%,%+81%3?417,+$P'&83%+'$`, , K:<,f17V,#,F3$%,%',0$'F,%+81%3?417,%',;'&&14'('$17", , 4=<,;O1,P'44'F+$5,%F',%&3+$7,413Q17,%'(36,3%,!a<aa38, 3$(,!a<-[38,P&'8,=%'*O3,.%3%+'$,%',;'&&14'('$17", =$6%O+$5,1471`, , K=<,K41371V,%144,81,%O1,/&+*1,'P,%O1,P+&7%,'$1", , 4F<,;O1,/&+*1,+$,?)7+$177,*4377,+7,-"H,1)&'7",=$6%O+$5, 1471`,, , KF<,I'V,%O3$0,6')", , 4O<,;O3$0,6'),d'7e",N3Q1,3,5''(,%&+/c,, , , 6,$=<&"<',-$%*17L*&7$%"

• 2-',"1$8*&7$%"7-"@%$H%"*%."/7-"

<,'3','%8'"*D$2&"8$%-21&7%+" &7='&*D1'-?" " 9/'"-;-&'="2-'-"&/'"82,,'%&" /$2,"*%."&/'".*&'"&$" <',-$%*17L'"&/'"&7='&*D1'-"

• ',078'?"9/','"7-"%$&"*%;"

7%3$,=*&7$%"*D$2&"$&/'," <$--7D1'"M2',7'-"<',3$,='."D;" &/'"2-',"7%"<,'07$2-".7*1$+-?" " 9/'"-;-&'="2-'-"&/'"&78@'&:81*--" <,'3','%8'-"7%812.'."7%"&/'"2-'," <,$371'?" " 6,$=<&"<',-$%*17L*&7$%" , ,

• +6/05&+$#/9)+$5&#0$%"@%&+$30(26%+0$,

:1*1$%,;&1$(7,

K*"(&$62"G&N$-O"P*,8Q*" D0#6%+0$G&A&$8/*">&*&7$%"

• 6%"&6$)&'+2"G&RSST:SU:SV5"T!SSA)"
• ".+#"G&6FAI"SS:VW:XV:YR:SZ:UT"

8JS"#%+."G&9$"@%$H"&7='&*D1'-"*%."<,78'-"&$"" " " """"""""9$,,'1$.$%'-"-&*&7$%, , , ,

gD7(<*'8/41D;6/1,$381#(1$%+P+*3%+'$, , gD7(<)71&h, , , gD7(<$381, , , gD7(<&'41K3771$51&, , , gD7(<4'*3%+'$.%3%+'$, , , gD7(<(3%1EaHE, , , gD7(<%+81, , , gD7(<A=Ci3((&177E!\EG!E9-EabE, , gjD7(<)71&h, gjD7(<*'8/41D;6/1h,

• +6/05&

4#"$6(+0, TID& E6#>65",

:1*1$%,;&1$(7,

I381<,d'7e,W3&*k3, EEEEEEEEEEEEEEEEEEE, W1$(1&<,A341,l,=51<,-^,l,J3$5)351<,./3$+7O,l.0+44,41Q14<,N+5O,l, K3%O'4'5+17<,I'$1,""", EEEEEEEEEEEEEEEEEEE, .177+'$7<, !-, l, =Q1&351i;)&$7<, H"[, l, =Q1&351i2)&3%+'$<, -"-, l, J37%i.177+'$<,-aa^EaGE-^,""", EEEEEEEEEEEEEEEEEEE, K&1P1&1$*17<,;+81%3?417V,;345',;&3+$V, b)7+$177,C4377V,=%'*O3,.%3%+'$V,!a<aa38""", EEEEEEEEEEEEEEEEEEE, C)&&1$%iJ'*3%+'$<,=%'*O3,.%3%+'$,l,K43%P'&8,T'$1,

K*"(&E(03+/",

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

=*])+&1,3,%'%34,'P,9aa,7)**177P)4,(+34'57,+$*4)(+$5,'&, $'%,*'$%1D%,+$P'&83%+'$,

@71&,7+8)43%+'$,%1*O$+])1", ;F',%6/17,'P,7*1$3&+'7<,'$1,'?M1*%+Q1,S.!>,'&,%F',S.->", C'8/3&1,?'%O,*'&/'&3,)7+$5,71Q1$,7%3%+7%+*34,8137)&17",

B37+$177,()&+$5,%O1,7+8)43%+'$", :1()*%+'$,+$,%O1,$)8?1&,'P,%)&$7", , ,

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

• D"&@M$,C,$'G,#+4?,*'S3"#$%&'#13'+4?1'$,1&+@*R'

' ' ' ' ' &

• N,%620$)4& $4& 0@,& "#,("5,& 46=C,(& )1& 06(4-& (,h6$(,%& 0)& 16*1$**& 0@,&

2)=/*,0,&-,0&)1&)CD,20$#,-&%,1$4,%&$4&0@,&-2,4"($)-;& &

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

• !"#$%&'*+)$,'

' ' ' ' ' & & &

• 7-,(-&4,,%&0)&,=/*)A&*,--&2)41$(="0$)4&06(4-&$4&0@,&1$4"*&"5,40;&
• N,%620$)4&$4&0@,&46=C,(&)1&-A-0,=&2)41$(="0$)4-&1)(&%"0"&$0,=-;&&
• +)0@& (,-6*0-& -@)?& 0@"0& 0@,& 1$4"*& 2)4#,(-"0$)4"*& "5,40& 2"(($,-& )60& "&

C,00,(&-,*,20$)4&)1&0@,&-A-0,=&(,-/)4-,-;&

:1*1$%,;&1$(7,

M0$%"@N6B6("&*7*%"2*&

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, F; !"#"$%&'("$)*G&?)61%6%+0$&

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, "H I9/%+20)6/&<$%"(6#%+0$&

G" =//4+*3%+'$,('83+$7, H" C'$*4)7+'$7,

;O1,5&13%,+$%1&17%,3$(,1D%1$7+'$,FO+*O,O3Q1,*)&&1$%46, &13*O1(,7'*+34,$1%F'&07,83017,/'77+?41<,

])+*046,+$%&'()*%+'$,'P,3,5&13%,$)8?1&,'P,3//4+*3%+'$7R, $1F,%6/17,'P,*'88)$+*3%+'$",

, K'77+?+4+%6,'P,+$%1&3*%+$5,+$,Q+&%)34,F'&4(7<,

/1&P1*%,7*1$3&+',%',(1Q14'/,8)4%+8'(34,767%187R, %301,+$%',3**')$%,$1F,+$P'&83%+'$,7')&*17R, 3(3/%,767%187,%',%O1,7/1*+P+*,)71&7,/&1P1&1$*17",

, #$%15&3%+'$,'P,*'$Q1&73%+'$34,351$%7,+$,Q+&%)34,F'&4(7<,

?'%,%O3%,/&'Q+(17,3*3(18+*,+$P'&83%+'$,+$,./3$+7O,+$,.1*'$(,J+P1",

,

:1*1$%,;&1$(7,

I9/%+20)6/&<$%"(6#%+0$&

.1*'$(, J+P1, S.J>, +7, 3, %O&11E(+81$7+'$34, Q+&%)34, F'&4(, 3**177+?41,'$,#$%1&$1%", , .1*'$(,J+P1,m+1F1&,344'F7,+%7,)71&7<,,

+$%1&3*%,F+%O,13*O,'%O1&,?6,813$7,'P,3Q3%3&7,F+%O,*3/3*+%6,'P, 8'Q181$%R, 3(Q3$*1(,41Q14,'P,7'*+34,$1%F'&0",

, (+PP1&1$%,%6/17,'P,*'88)$+*3%+'$,38'$5,&17+(1$%7,,

517%)&17,S3$+83%+'$7>V, %1D%,81773517,S*O3%>V, Q'+*1",

:1*1$%,;&1$(7,

I9/%+20)6/&<$%"(6#%+0$&

.J, +7, *)&&1$%46, )71(, F+%O, 7)**177, 37, 3$, 1()*3%+'$34, /43%P'&8,+$,83$6,+$7%+%)%+'$7<,

+%, +7, '$1, 'P,%O1, 8'7%, /'/)43&, Q+&%)34, 7'*+34, F'&4(7, S8+44+'$7, 'P, )71&7>R,, +%, )717, Q1&6, 3(Q3$*1(, %1*O$'4'5+17, P'&, %O1, (1Q14'/81$%, 'P, &134+7%+*,7+8)43%+'$7, 83$6,/'77+?+4+%+17,P'&,*)7%'8+T3%+'$R,, 1$*')&3517,+$$'Q3%+'$",

:1*1$%,;&1$(7,

I9/%+20)6/&<$%"(6#%+0$&

_1, O3Q1, (1Q14'/1(, 3, *'$Q1&73%+'$34, 81%3?'%, %O3%, P3*+4+%3%17, 3*3(18+*, +$P'&83%+'$, S7)?M1*%7V, 41*%)&1&7V, ('*%'&34, 7%)(+17V, 3$(, (13(4+$17>,?371(,'$,%O1,P13%)&17,/&'Q+(1(,?6,3,/&1Q+')746,(1Q14'/1(, *'$Q1&73%+'$34,351$%", ,

:1*1$%,;&1$(7,

I9/%+20)6/&<$%"(6#%+0$&

!"#$%&'()*+,&%'*-%' .,/*)#"'0(/"&1' 2(,$%'#3&'0%4' +%/.%/+'' 5%$-3("(6,%+7' 89:;<5='>?%%$-@,4=' >@2(,$%='2,.(A=' B,$/(+(C*'>D!E=' @(F)%3&('55>:''

!"#$%&'()*"#(+,(-%#).,

:1*1$%,;&1$(7,

I9/%+20)6/&<$%"(6#%+0$&

/&01*)%0)2&%,)",3%$%+"4,(,0"#$%&'()*"#(+,(-%#),*#,56,,

:1*1$%,;&1$(7,

K&14+8+$3&6,1Q34)3%+'$,?6,813$7,'P,%O1,3*])+7+%+'$,'P,Ha,(+34'57, , , , , , , .+8+43&,?1O3Q+')&,%',%O1,*'$Q1&73%+'$34,351$%,)7+$5,'$46,Q'+*1, , :1])+&1(<, %', 1Q34)3%1, $1F, P13%)&17, %', ?1, +$*4)(1(, +$, %O1, *'$Q1&73%+'$34, 81%3?'%,%',+8/&'Q1,%O1,*'88)$+*3%+'$,/&'*177", %', *3&&6, ')%, 3, (1%3+41(, 3$3467+7, 'P, %O1, +$%15&3%+'$, 'P, (+PP1&1$%, 8'(34+%+17",

:1*1$%,;&1$(7,

I9/%+20)6/&<$%"(6#%+0$&

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, 9" :1*1$%,;&1$(7<,=(3/%3%+'$,

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

O; ?11/+#6%+0$&)026+$*& H" C'$*4)7+'$7,

=//4+*3%+'$,2'83+$7,

Provide Information: Conversational agent providing information in natural language about train services, schedules, and fares in Spanish.

=//4+*3%+'$,2'83+$7,

Carry out a specific task: Development of a conversational agent for the booking of sport facilities.

!3:TI*)5&$-&"&2)4#,(-"0$)4"*&"5,40&?@$2@&"20-&"-&"&#)$2,&*)5C))H&0)&2)**,20& @)=,&=)4$0)(,%&%"0"&1()=&/"0$,40-&-611,($45&1()=&KA/,Id&%$"C,0,-;& & K@,&%"0"&2)**,20,%&CA&0@,&-A-0,=&"(,&0@,&/"0$,40U-&?,$5@0F&C*))%&/(,--6(,& O-A-0)*$2&"4%&%$"-0)*$2&#"*6,-P&"4%&-65"(&*,#,*-;&& & K@,&-A-0,=&#"*$%"0,-&"4%&"4"*A-,-&0@,&%"0"E& /()#$%$45& -)=,& $==,%$"0,& 1,,%C"2H& 0)& 0@,& /"0$,40-& (,5"(%$45& 0@,$(& 26((,40&/()5(,--k& 2)==64$2"0$45& 0@,& (,-6*0-& 0)& %)20)(-& "0& 0@,& @)-/$0"*& ?@)& "(,& "C*,& 0)& (,#$,?& 0@,& /"0$,40U-& /()5(,--& 5("/@$2"**A& "4%& %,"*& ?$0@& "4A& "*,(0-& 5,4,("0,%&CA&0@,&-A-0,=&2)42,(4$45&"C4)(="*&%,#,*)/=,40-;& &

&

E-Health:

=//4+*3%+'$,2'83+$7,

E-Health:

=//4+*3%+'$,2'83+$7,

K"H$45&$40)&"22)640&0@,&(,h6$(,=,40-& )1& 0@,&0"-HF&"4&$4$0$"*&-0("0,5A&?"-& %,-$54,%&1)(&0@,&!9E& & &

& Identification and waiting phase The system analyzes which data is required for the current user and requires it. Every item is confirmed after the user has provided its value. The user can only provide

• ne item at a time.

System analysis and diagnosis

E-Health:

=//4+*3%+'$,2'83+$7,

=//4+*3%+'$,2'83+$7,

Carry out a specific task: Learning foreign languages

=//4+*3%+'$,2'83+$7,

Access for disabled people: Voice Google

=//4+*3%+'$,2'83+$7,

Interaction in virtual worlds: Design of a chatbot to provide academic information

2(,$%'!(/*#"+

=//4+*3%+'$,2'83+$7,

G>>'3%0+'+?%%$-H4#+%&'+%/.,$% I",6-*'>,J)"#*,(3'D++,+*#3*

=//4+*3%+'$,2'83+$7,

>*)&%3*+'K),&#3$%'L/%$(JJ%3&,36' +M+*%J+N' O3H",3%'+)/.%M+

=//4+*3%+'$,2'83+$7,

5/#.%"H?"#33,36'+M+*%J B(.,%'/%$(JJ%3&#*,(3'+M+*%J >?%%$-'E3*%/C#$%'C(/' K((6"%B#?+

=//4+*3%+'$,2'83+$7,

89%/+$"& !" #$%&'()*%+'$,

• "

./'01$,2+34'5,.67%187, 9" :1*1$%,;&1$(7<,=(3/%3%+'$,

3> .%3%+7%+*34,(+34'5,83$35181$%, ?> @71&,A'(14+$5, *> B8'%+'$,:1*'5$+%+'$, (> C'$%1D%E3F3&1,767%18, 1> A)4%+8'(34,#$%1&3*%+'$,

G" =//4+*3%+'$,('83+$7, P; M0$#/9*+0$*&

C'$*4)7+'$7,

• Main research lines in the Applied Artificial Intelligence Group:

Development of dialog systems for specific tasks; Integration of dialog systems in multiagent systems; Context-aware systems; Multimodal applications (speech+images+video+text, interaction

in virtual worlds);

Affective computing and HCI. Statistical dialog methodologies for dialog management; User modeling and simulation; Design of user-centered applications and recommending

systems.