[PPT] - Forest-to-String SMT for Asian Language Translation: NAIST at WAT PowerPoint Presentation

SLIDE 1

1

NAIST at WAT 2014

Forest-to-String SMT for Asian Language Translation: NAIST at WAT 2014

Graham Neubig Nara Institute of Science and Technology (NAIST) 2014-10-4

SLIDE 2

2

NAIST at WAT 2014

Features of ASPEC

Translation between languages with different

grammatical structures

流動プラズマを正確に測定するために画像を再構成した。 an image was reconstituted in order to measure flowing plasma correctly .

We all know: Phrase-based MT is not enough

for the accurate measurement of plasma flow image was reconstructed .

SLIDE 3

3

NAIST at WAT 2014

Solution?: 2-step Translation Process

Pre-ordering [Weblio, SAS_MT, NII, TMU, NICT]
RBMT+Statistical Post Editing [TOSHIBA, EIWA]

我々は科学論文を翻訳する我々翻訳する科学論文 we translate scientific papers 我々は科学論文を翻訳する we translate science thesis we translate scientific papers

SLIDE 4

4

NAIST at WAT 2014

This is a lot of work... :(

How do I make good Japanese-English preordering rules?! How do I make good Japanese-Chinese preorderering rules?! What about error propagation? What if better preordering accuracy doesn't equal better translation accuracy?

SLIDE 5

5

NAIST at WAT 2014

Evidence

SLIDE 6

6

NAIST at WAT 2014

Our Solution: Tree-to-String Translation [Liu+ 06]

友達とご飯を食べた SUF5 VP0-5 PP0-1 VP2-5 PP2-3 N2 P3 V4 N0 P1 VP4-5 a meal

a meal

x1 x0 x1 x0 my friend

my friend

x1 with x0 x1 x0

with

ate

SLIDE 7

7

NAIST at WAT 2014

Requirements for a Tree-to-String Model

This is a test . It uses data .

これはテストです。データを使用します。

Parallel Corpus

Source Sentence Parser

Alignments

Rule Extraction Rule Scoring Optimization

Tree-to-String Model

SLIDE 8

8

NAIST at WAT 2014

Reducing our work load.

How do I make good Japanese-English preordering rules?! How do I make good Japanese-Chinese preorderering rules?! What about error propagation? What if better preordering accuracy doesn't equal better translation accuracy?

X X X

SLIDE 9

9

NAIST at WAT 2014

Forest-to-string Translation [Mi+ 08]

I saw a girl with a telescope

PRP 0,1 VBD 1,2 DT 2,3 NN 3,4 IN 4,5 DT 5,6 NN 6,7 NP 5,7 NP 2,4 PP 4,7 VP 1,7 S 0,7 NP 2,7 NP 0,1

SLIDE 10

10

NAIST at WAT 2014

Travatar Toolkit

Forest-to-string translation toolkit
Supports training, decoding
Includes preprocessing scripts for parsing, etc.
Many other features (optimization, Hiero, etc...)

Available open source! http://phontron.com/travatar

SLIDE 11

11

NAIST at WAT 2014

NAIST WAT System

SLIDE 12

12

NAIST at WAT 2014

WAT Results

en-ja ja-en zh-ja ja-zh 10 20 30 40 50 BLEU en-ja ja-en zh-ja ja-zh 20 40 60 HUMAN Other NAIST

First place in all tasks!

+2.2 +2.7 +3.6 +1.8 +13.0 +15.0 +28.3 +3.8

SLIDE 13

13

NAIST at WAT 2014

System Elements

Travatar!

Same as [Neubig & Duh, ACL2014]

Recurrent Neural Net Language Model

Pre/post Processing (UNK splitting, transliteration) Dictionaries

SLIDE 14

14

NAIST at WAT 2014

Recurrent Neural Network LM

Vector representation → robustness
Recurrent architecture → longer context

I can eat an apple </s>

SLIDE 15

15

NAIST at WAT 2014

Pre/post processing

UNK segmentation (ja-en)

球内部球内部試験管立て試験管立て

Kanji Normalization (ja-zh, zh-ja)

イチョウ黄叶イチョウ黄葉臭気鉴定师臭気鑑定師

Transliteration (ja-en)

Japan インテック Japan Intekku Dictionary addition (ja-en) 膿瘍 apostema 典型 archetype

SLIDE 16

16

NAIST at WAT 2014

Conclusion

SLIDE 17

17

NAIST at WAT 2014

Future Work

LOSE at next year's WAT.

(Make Travatar so easy to use that others can use it to make really good MT systems for Asian languages.)

Starting soon! Training scripts to be available: http://phontron.com/project/wat2014