Transforming IbnSina Into an Advanced Multilingual Interactive ...

2 downloads 0 Views 131KB Size Report
application in malls, schools, as a receptionist etc. ,QGH[ 7HUPV² Intelligent Robots, Android,. Multilingual, Interactive, Natural Language Interfaces. 1.
2011 IEEE GCC Conference and Exhibition (GCC), February 19-22, 2011, Dubai, United Arab Emirates

75$16)250,1*,%16,1$,172$1$'9$1&(' 08/7,/,1*8$/,17(5$&7,9($1'52,'52%27

 1LNRODRV0DYULGLV$OLD$O'KDKHUL/DWLID$O'KDKHUL0DLWKD.KDQML1RXUD$O'DUPDNL  ,QWHUDFWLYH5RERWVDQG0HGLD/DE8QLWHG$UDE(PLUDWHV8QLYHUVLW\ $%675$&7



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² ,QWHOOLJHQW 5RERWV $QGURLG 0XOWLOLQJXDO,QWHUDFWLYH1DWXUDO/DQJXDJH,QWHUIDFHV  

 ,1752'8&7,21

,EQ6LQD >@ LV WKH ZRUOG¶V ILUVW $UDELFODQJXDJH FRQYHUVDWLRQDO DQGURLG URERW ZKLFK DOUHDG\ KDV FDSDELOLWLHVIRUJHQHUDWLQJSUHVFULSWHGIDFLDOH[SUHVVLRQV DVZHOODVKHDGPRYHPHQWVDQGIRUXWLOL]LQJ$UDELFWH[W WRVSHHFK DQG VSHHFK UHFRJQLWLRQ IRU WKH FUHDWLRQ RI VLPSOH GLDORJXHV DV GHVFULEHG LQ >@>@ )XUWKHUPRUH ,EQ6LQD LV SDUW RI DQ LQWHUDFWLYH WKHDWUH ZLWK PXOWLSOH SRVVLELOLWLHV IRU KXPDQ WHOHSDUWLFLSDWLRQ WKURXJK PRWLRQ FDSWXUH YLUWXDO ZRUOGV HWF >@ >@ +RZHYHU LQWHUHVWLQJ WKHFXUUHQWUHVXOWVPLJKWEHWKHUHDUHVWLOOYHU\PDQ\RSHQ SRVVLELOLWLHVIRUH[WHQVLRQ 7KH SXUSRVH RI WKLV SURMHFW LV WR H[WHQG ,EQ6LQD¶V FDSDELOLWLHV LQ PXOWLSOH ZD\V VR WKDW LW FDQ EHFRPH DQ H[FLWLQJHGXFDWLRQDOSHUVXDVLYHURERWLQWKHIXWXUH7KH PDLQ D[LV IRU H[WHQVLRQ ZHUH DFFHVV WR RQOLQH FRQWHQW :LNLSHGLD  DQG VWRUHG FRQWHQW .RUDQ GDWDEDVH  IRU GLDORJXH JHQHUDWLRQ EDVLF PXOWLOLQJXDO FDSDELOLW\ H[SORUDWLRQ (QJOLVK DQG $UDELF DOVR XWLOL]LQJ *RRJOH 7UDQVODWH  EDVLF UHDGDORXGZULWWHQWH[W FDSDELOLW\

978-1-61284-119-9/11/$26.00 ©2011 IEEE

120

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¶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

 

 )LJXUH0RGXODU6XEV\VWHPVRI,EQ6LQD

6 %DVLF5HDGLQJ6XEV\VWHP ,QLPSOHPHQWLQJWKH2&5ZHILUVWFRQQHFWHG,EQ6LQDZLWK WKH FRQWURO FRPSXWHU ,EQ6LQD¶V H\H FDPHUDV SURYLGH XV ZLWK YLGHR VR ZH FDQ SXW D SLHFH RI SDSHU RI ZKLWH EDFNJURXQGZLWKEODFNWH[WLQIURQWRIKLPWREHUHDGRXW 7KHFRGHWDNHVDVFUHHQVKRWIURPWKHFRPSXWHUIURPZKDW ,EQ6LQD VHHV WKURXJK FDPHUD FDSWXUH 7KH VFUHHQVKRW LV WKHQ FRQYHUWHG IURP MSJ WR SQP 7KH2&5FRGHRXWSXWV WKH WH[W LQ WKLV SLFWXUH DQG VDYHV LW LQ D WH[W GRFXPHQW WKURXJK WKH 2&5 PRGXOH >@ $FDSHOD WH[WWRVSHHFK VSHDNVXSWKHWH[WIURPWKHRXWSXWRIWKH2&5FRGH

122



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³IUR]HQ´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

123

WRDQDWWUDFWLYHURERWWKDWFDQILQGUHDOZRUOGDSSOLFDWLRQ LQ PDOOV VFKRROV DV D UHFHSWLRQLVW HWF $QG PRVW VLJQLILFDQWO\ WKLV ZRUN SURYLGHV LPSRUWDQW UHVHDUFK LQVLJKWV DQG D ZRUNLQJ UHDOZRUOG SURRIRIFRQFHSW RI FRUSXV DV ZHOO DV RQOLQHLQIR PLQLQJ FRQYHUVDWLRQDO DQGURLGVZLWKPXOWLOLQJXDODVZHOODVUHDGLQJDQGGLVSOD\ FDSDELOLWLHV DQG ZKLFK LQ WKH IXWXUH PLJKW ZHOO VXSSRUW PXOWLSOH FRQYHUVDWLRQDO SDUWQHUV DV ZHOO DV VXVWDLQDEOH ORQJHUWHUP UHODWLRQV ZLWK KXPDQV DQG WKXV HYHQWXDOO\ EHFRPHLPSRUWDQWGDLO\DVVLVWDQWVDQGIULHQGVWRXV 

 

 5()(5(1&(6

>@ 1 0DYULGLV DQG ' +DQVRQ 7KH ,EQ6LQD &HQWHU $Q $XJPHQWHG5HDOLW\7KHDWHUZLWK,QWHOOLJHQW5RERWLFDQG9LUWXDO &KDUDFWHUV,(((5R0$1 >@ $KPHG 'DKHUL 6DHHG 0HUUL 0XKDPHG .XZDLWL 6DOHP .DWKHUL &RQYHUVLQJ ZLWK ,EQ 6LQD 7RZDUGV 'LDORJLF 5RERWV ZLWK $UDELF /DQJXDJH *HVWXUDO DQG ([SUHVVLRQ &DSDELOLWLHV 6HQLRU3URMHFW5HSRUW8$(8&,7 >@ /5LHN10DYULGLVHWDO³,EQ6LQD6WHSV2XW([SORULQJ $UDELF$WWLWXGHV7RZDUG+XPDQRLG5RERWV´$,6% >@ 10DYULGLV(0DFKDGRHWDO5HDOWLPH7HOHRSHUDWLRQ RIDQ,QGXVWULDO5RERWLF$UP7KURXJK+XPDQ$UP0RYHPHQW ,PLWDWLRQ,5,6 >@ & &KULVWRIRURX 1 0DYULGLV HW DO$QGURLG WHOH RSHUDWLRQ WKURXJK %UDLQ&RPSXWHU ,QWHUIDFLQJ $ UHDOZRUOG GHPRZLWKQRQH[SHUWXVHUV,5,6 >@ ³/DWHQW 6HPDQWLF $QDO\VLV SUREOHP´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´-DYD+RZWR3URJUDP´3UHQWLFH+DOO >@ ³7XWRULDO RQ PDNLQJ DQ $UWLILFLDO ,QWHOOLJHQFH &KDWERW´ $YDLODEOHIURPDLSURJUDPPLQJFRPMDYDBERWBWXWRULDOKWP >@ $KPDV+DPPDG-DYDDQG$UDELF6XSSRUW'HF DKPEORJVSRWFRPMDYDDQGDUDELFVXSSRUWKWP >@ ©-RHUJ 6FKXOHQEXUJ ³RSHQVRXUFHFKDUDFWHUUHFRJQLWLRQ´ 5HWULHYHGIURPMRFUVRXUFHIRUJHQHWLQGH[KWPO >@ 66&6HUYR&RQWUROOHU%RDUG+DUGZDUH0DQXDO >@ 52%27,6'\QDPL[HO5;0RWRU+DUGZDUH0DQXDO >@ 7.DQXQJR*0DUWRQ2%XOEXO³2PQL3DJHYV6DNKU 3DLUHG 0RGHO (YDOXDWLRQ RI 7ZR $UDELF 2&5 3URGXFWV´ FLWHVHHU[LVWSVXHGXYLHZGRFGRZQORDG"GRL  U HS UHS W\SH SGI >@ % - )RJJ ³3HUVXDVLYH WHFKQRORJ\ XVLQJ FRPSXWHUV WR FKDQJHZKDWZHWKLQNDQGGR´0RUJDQ.DXIPDQQ >@ 1 0DYULGLV : .D]PL 3 7RXOLV & %HQ$EGHO.DGHU 2Q WKH V\QHUJLHV EHWZHHQ RQOLQH VRFLDO QHWZRUNLQJ )DFH 5HFRJQLWLRQDQG,QWHUDFWLYH5RERWLFV&D6R1