• No results found

Interactive cross-language document selection

N/A
N/A
Protected

Academic year: 2019

Share "Interactive cross-language document selection"

Copied!
38
0
0

Loading.... (view fulltext now)

Full text

(1)

This is a repository copy of

Interactive cross-language document selection

.

White Rose Research Online URL for this paper:

http://eprints.whiterose.ac.uk/4560/

Article:

Oard, D., Gonzalo, J., Sanderson, M. et al. (2 more authors) (2004) Interactive

cross-language document selection. Information Retrieval, 7 (1-2). pp. 205-228. ISSN

1573-7659

https://doi.org/10.1023/B:INRT.0000009446.22036.e3

[email protected]

https://eprints.whiterose.ac.uk/

Reuse

Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright

exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy

solely for the purpose of non-commercial research or private study within the limits of fair dealing. The

publisher or other rights-holder may allow further reproduction and re-use of this version - refer to the White

Rose Research Online record for this item. Where records identify the publisher as the copyright holder,

users can verify any specific terms of use on the publisher’s website.

Takedown

If you consider content in White Rose Research Online to be in breach of UK law, please notify us by

(2)

promoting access to White Rose research papers

White Rose Research Online

Universities of Leeds, Sheffield and York

http://eprints.whiterose.ac.uk/

White Rose Research Online URL for this paper:

http://eprints.whiterose.ac.uk/4560/

Published paper

Oard, D., Gonzalo, J., Sanderson, M., Lopez-Ostenero, F. and Wang, J. (2004)

Interactive cross-language document selection. Information Retrieval, 7 (1-2).

pp. 205-228.

(3)

£

Ý

Þ

Ý

£

!

""#

$ %

&

£

!"# $%&'(**

Ý

& + ,-+

-*.** / 0!#!,$&12(**

Þ

3Æ 4 53Æ 6! .,7

(4)

!

"

#

$ %& $ &

' %

$(& %

¯ $&%

¯ %

!

%

)* +,

$*+& $-" . //0℄& "

(5)

$ &

¯

¯

! 3 ! %

¯ $ &

4

5

!

% !

%

4 !

3 3

!

6

3

)7 , $7& $ &

8

7

7 ' % %

(6)

#9 #:

#;

8 *+

( 8

'

*+

* "

*+

% <

% $

%

-" . //0℄& +

=

+ %

-6 //0℄

%

(7)

% 8

% >

?

% # %

% 8

+ * $+(*& @#A BA*##4B *

+ $B*+& C * (

5 $*(5& (

), )

, -* * 2DDD℄

A

$

&

=

7

(8)

8 > ),

8 > $ &

>)8,

% $ &

A 8 >

!

F

6 %

" 3 $ &

$2& $9& 7 7 $7&

$ &

),

A3

7 3

%

% 3

%

$

&$& $&

# % $ 7

(9)

73

$ & $2&

$9& '8

$ 3 &

A

),

3

8

'

$ -6 ///℄&

%

'

% % G

3 5

'

G %

%

(10)

5 %

(

*%

Æ

?

+ * $+(*& //:

- " //0℄ //; //H

+(* 2DD2

>

?

*+ #

¯ +

-" + /// + //H℄ 6

( C

$ & ? 6

? 3

8 '

$ &

- ' //0 ' //0℄

(11)

G % (

%

+AB#>

¯ ( 7@B(I ? %%

' *+

# 7 -* ///℄ A

%

¯ ". %

-" ///" . 2DDD℄

#J22

J

//K 0;K

C ? +(*

7 %

3

%

3

¯ #3 %

(12)

;D ? ?

?

?

A

$

& '

'

' ?%

% % ?

?$

(13)

< # 5

( '

(

5

¯ ( 5 #

*+ ! %

¯ A ED *+

@ 3

' 5 % C 6 @

A <( ( % @

7 5 $ DDD&

ED

? ED

¯ 5 *(52DDD

¯ A ( # <

9D 7 $7& #

5 ),

? ),

' :D*(5 2DDD

(14)

¯ . ?

$ &

¯ 5 ?

3

¯ 5 ED

$& 9; 2H

9$& ;

H$& 2 9: ; H H 2/

2/$& 2099 2D29 20

# $& $&

ED

"

*(52DDD

), %) ,

(15)

$ ED & 99

$&

), ) , ) ,

*(5 2DDD )@, )

?,

(

% A 2D

)

% -℄

, < %

4

' ?

<

% +(*

A

2

(16)

# H # 2 92/

2 # 2 H # 92/

9 # H # 2 2/ 9

: # 2 H # 2/ 9

E # H # 2 2/ 9

; # 2 H # 2/ 9

H # H # 2 92/

0 # 2 H # 92/

2 < L 9 H L2/

:

A M+?!

« N

O$ &

-+? /H/℄

N $

¾

O & N 2D N D2 8

# N DE N D0 8

½

5

N DE

?

-" 2DD ℄ *(5

? $&?

½

8

¬

9

¬

(17)

5

$

& ' ) ,

) ,

: 5 0

' $& '

-< > 2DDD℄

4 4

$AB"MA& DDE

% %

7 @B(.

P NP O O

' @ B (Q .

$@B(.&# @7$@7.& @#

@ #Æ $#6(5& @ R

(18)

#

@B(. # (

5

@ #Æ( (

( 5

!

A @ 7

'

$ ),& # 7

( :D !

=

' 5

5 7

(

'

9EDDD (5

A

=

5 5

=

5 ?

7

¾

3 : 3 ;+2< * !!6

(19)

# $ %

' 3

# ' ?

A

;

B

' S2D

(20)

# 7 J"## 7 J"## 7 J"##

D D;2 D20 DD DH0 D0 DE9

D2 D9: D 9 DH0 DDD DE; DDH

D9 D 9 D D DD DDD DE2 DDE

D: D 9 D2H D/ D09 DE2 DEE

A D9 D2D D/2 D: D; D2/

9 &

$ %

7

7 AAB"MA ; 7

$ DD2& ?

7

9

# ?

:DK 3

?

ED

D2; A

7

% 5

7

( % ?

(21)

" " $ ) $ ! "

%

),? ) , ?

%

$ & ) ?, "

3

A 9/0 ) , ?

$ & 2D ), ? $

& ' ?

?

A A

?

Æ A ?

7 %$

& Æ

(22)

!

% @B(. #

) ,

"

% % 8

$

&

Æ ?

7 5

?

' ! "

-<P 2DD ℄ 5(

5

# ' 2;HDDDDD #

*(52DDD # )(5(, 2EDDDD

//: " 9;DDDDD

#

# # # '

(

(23)

'

(

$ &

(5(

$ %$&$

%

$

2 * $ " +

# '

$

3&

(

(

=

% $ &

59

'

(24)

( 5

( 7

$

&

! $ &

$

& :

# (

E2K # (

5

(25)

# 7 <6+A#( 7 <6+A#( 7 <6+A#(

D DD/ DDD D:0 DDD D2/ DDD

D2 D2E DE9 D09 DHD DE: D;2

D9 D9D D90 D/D D2H D;D D99

D: DD0 D90 DDD D9E DD: D9H

DE DDD D / DDD DDD DDD D D

D; DDD D:2 D 0 DE0 DD/ DED

DH D90 DDD D09 D9E D; D 0

D0 DD9 D02 D2H DH D E DHH

A D : D9: D:: D9H D2/ D9;

: ,-.&

$ %

+

7

?

A AB"MA 92 (

$ND2D& $ND :& ?

#DE

5 % DE

? %

(26)

@7@B(.

7

7

? Æ

%

%

%

3 !

7=

# $

&

" ! # !

#Æ ( (

( #Æ

(( #

5 A

5 5

' 2D $S9D&

@

(27)

% # $2:D

( HD5&

*(5 ?

" ?

¯ 3

5

# 5 7

¯ 3 *(5

5

*(5$&

(

#Æ E

?

? A

!

?

" ?

!

?

3 3

=

(28)

# 7 7"B" 7 7"B" 7 7"B"

D DH9 D9D D09 D20 DH: D9D

D2 D:E D;9 D/ D09 DE9 D;H

D9 DE; D9; DDD DDD D:H D9E

D: D;: D9 D/ DH D;0 D:D

DE DHE D9H DD D9; DHH D9H

D; DEE D:: DH/ DE; D; D:;

DH D:2 D;9 DDD DE; D: D;

D0 D / D90 D/ DH D9/ D:E

A DE: D:9 D;H DED DE0 D:E

E Æ&

$ %

+(*

ED -M //0℄ #

-# //0℄

" ED %

M '

#Æ *(5

D9/ D:H $ ) , ? ),&

$ &

? $ ?

? 2: &= $2& *(5

(29)

0

20

40

60

80

100

0

20

40

60

80

100

Precision

Recall

UMD-MT

UMD-Gloss

SHEF-MT

5: " 5

) , ), ?= $9&

F

*(5

= $:& *(5 Æ ?

8

F = $E& *(5

$ &

5: 5

5 E ( A

D9D

( D22 5 A

#Æ 7 7

5 $

(30)

0

20

40

60

80

100

0

20

40

60

80

100

Precision

Recall

SHEF-Monolingual

UNED-MT

UNED-Phrases

5 E " (

7 7

8

7

$ % &

' $

&

(31)

#6(57 E/ :D :E 9/

@B(.< :H 9: 9E 92

@B(.7 :0 22 20 2

5

@7.7 H; E0 ; EH

#6(57 ;H :; E/ :0

@7.J E 2H 2/ 2;

; *

"

3

$

& 3

; 5:

E A

H $

:& A

$ &

5

$ '( ) *

' ), )

, '

(32)

# 7 <6+A#( 7 <6+A#( 7 <6+A#(

D DD2 DDD D / DDD D DDD

D2 D:D D22 DEE D/D D:0 DE;

D9 DD/ D;/ DHD D: D:D DEE

D: DD; D 9 DDD D:E DD9 D2/

DE DDD D 0 DDD DDD DDD DD/

D; DDD D E D9E D;9 D H D9/

DH D 9 DDD DEE D:E D9: D29

D0 DDE D;D D: D90 D29 D:/

A D ; D2E D9: D:D D22 D99

$

& D : D9: D:: D9H D2/ D9;

H ,-.&

(33)

/

# D; D2/ D9; D2/ D;D D:;

0 D;H D:2 D90 D:D DE/ DE2

0 * ?

$ & *(5

3

), ) ,

0

$&

7 $

& ? '

-' " 2DD ℄ 5

)

, ),

* *(5

1% A $

&

(34)

* (

5

2 % *(5 2DD2

2DD9

7

* ?

" *(5 %

¯ +?

5*(5 2DD2

% %

5

% ?

3

¯ " )

, ),

'

? *(5 2DD2 2DD9

¯ 7

$3 &

(35)

2DD9 2DD9

" 5

?

% *

? $4 4 ?

&

+

"

> %

%

+ A TU > * *3 > 6 J

< 7B 5QV7 < < " * < .

< .A+<A

B;;DD DD20/ D$.(#&(@ ?#2DDD2E9 D$*& #2DDD

(36)

-* ///℄ * C . A R ( J @3 6

A 7 $ ///& A

& " 9;$2&2HEW20/

-* * 2DDD℄ * . > * A . $2DDD& & '

&( @(*"

-6 ///℄ 67 A$ ///& @ 3 >3X +

+B> " & DA'

B X 44444 4 D

-6 //0℄ 6 ' A < # * > R . #

" . $ //0& . Y

)*& " &+&

# & HW2:

- " //0℄ ( " < $ //0& *

+(*;

), & " &+&

# &

-Q 3" 2DD ℄ Q 3" 5 J3C <P A M? 5

$2DD & B

-# '. / 0 ./')11,

-7 ///℄ 7 # # ( 5 B $ ///& #

## &'

(37)

& ! 99

A# #

-" 2DD ℄ " . ' JA *3 * $2DD & *(5

7 # < *

/ '. /

-" + ///℄ " . ' + <$ ///& #

&

"9E$9&9;9W9H/

-" ///℄ " ' * C . 7 ( 7# 6

#6 $ ///& J A

3 2 " .3 &+&

-# " &

-" . 2DDD℄ " ' * . 7 ' $2DDD&

** 4

&

-<P 2DD ℄ <P A J3C M? 5 $2DD & *

## . & '

B 2 W 9D

-< > 2DDD℄ < C > . $2DDD& "('5

'. #

-+ //H℄ + <$ //H& ( ' &

# '. !( # A A

(38)

6 & &

7 "

-# > 2DD ℄ # 7 > T $2DD & *(5 #Æ

-# '. / 0./')11,

-#3 2DD ℄ #3 7 B 6 R $2DD & A

# 4 9E$:&:2 W:90

- ' //0℄ R ' C $ //0& < 7

@ ? 5 . J 6 (

! " !

9;:W9H9 # B A E2/

- +? /H/℄ +?* C$ /H/& & >

-M //0℄ M($ //0& M ?

), & "

&+& # &

-' " 2DD ℄ ' C " . ' $2DD & *(5 2DD 7

* -# '

. / 0./')11,

-' //0℄ 'C#R>$ //0&A

/ & .

References

Related documents

Our observations of the ra- dio continuum emission from IRAS 05373 + 2349 VLA 2 indicate that it arises in a jet, similar to those observed towards low- and high-mass young stars

(I I) to use the transitional period to draw up a report on the conditions under which financial transactions are taxed or exempted In order to examine the

the Council also examined a report from the Permanent Representatives Committee. submitting a number of questions which are fundamental for the Iiberalization

The developed cell model is implemented in DEM by using a clump of nine particles, representing atrial cellular geometry; (ii) a 2D DEM tissue model of atrial tissue was developed

While drawing encouragement from the largely positive commentaries on my paper, I then take on three main critiques of the paper – first that it has inadvertently promoted

Figure 2 Influence of single amino acid substitutions on binding of human monoclonal antibodies and patients' sera to IA-2. A: Influence of single amino acid substitutions on binding

In its Opinion, adopted unanimously with I abstention, the Commit- tee endorses the Commission proposal to boost NCI resources as a supplementary means of

A recent Dutch case brought to the attention of the European Parliament and of the European Commission is probably a typical example repeated elsewhere in