Upload
bena
View
32
Download
0
Tags:
Embed Size (px)
DESCRIPTION
The AVENUE Project: Bootstrapping MT Prototypes for Languages with Limited Resources. Faculty: Alon Lavie , Jaime Carbonell, Lori Levin, Ralf Brown Students: Erik Peterson, Christian Monson, Ariadna Font-Llitjos, Alison Alvarez. Progression of MT. Started with rule-based systems - PowerPoint PPT Presentation
Citation preview
The AVENUE Project:Bootstrapping MT Prototypes for Languages with Limited
Resources
Faculty: Alon Lavie, Jaime Carbonell, Lori Levin,
Ralf Brown
Students:Erik Peterson, Christian Monson,
Ariadna Font-Llitjos, Alison Alvarez
September 7, 2005 AVENUE Project 2
Progression of MT• Started with rule-based systems
– Very large expert human effort to construct language-specific resources (grammars, lexicons)
– High-quality MT extremely expensive only for handful of language pairs
• Along came EBMT and then SMT…– Replaced human effort with extremely large volumes of
parallel text data– Less expensive, but still only feasible for a small number of
language pairs– We “traded” human labor with data
• Where does this take us in 5-10 years?– Large parallel corpora for maybe 25-50 language pairs
• What about all the other languages?• Is all this data (with very shallow representation of
language structure) really necessary?• Can we build MT approaches that learn deeper levels of
language structure and how they map from one language to another?
September 7, 2005 AVENUE Project 3
Why Machine Translation for Languages with Limited Resources?
• We are in the age of information explosion– The internet+web+Google anyone can get the
information they want anytime…• But what about the text in all those other
languages?– How do they read all this English stuff?– How do we read all the stuff that they put online?
• MT for these languages would Enable:– Better government access to native indigenous and
minority communities– Better minority and native community participation in
information-rich activities (health care, education, government) without giving up their languages.
– Civilian and military applications (disaster relief)– Language preservation
September 7, 2005 AVENUE Project 4
The Roadmap to Learning-based MT
• Automatic acquisition of necessary language resources and knowledge using machine learning methodologies:– Learning morphology (analysis/generation)– Rapid acquisition of broad coverage word-to-word
and phrase-to-phrase translation lexicons– Learning of syntactic structural mappings
• Tree-to-tree structure transformations [Knight et al], [Eisner], [Melamed] require parse trees for both languages
• Learning syntactic transfer rules with resources (grammar, parses) for just one of the two languages
– Automatic rule refinement and/or post-editing• Effective integration of acquired knowledge
with statistical/distributional information
September 7, 2005 AVENUE Project 5
CMU’s AVENUE Approach• Elicitation: use bilingual native informants to produce a
small high-quality word-aligned bilingual corpus of translated phrases and sentences– Building Elicitation corpora from feature structures– Feature Detection and Navigation
• Transfer-rule Learning: apply ML-based methods to automatically acquire syntactic transfer rules for translation between the two languages– Learn from major language to minor language– Translate from minor language to major language
• XFER + Decoder:– XFER engine produces a lattice of all possible transferred
structures at all levels– Decoder searches and selects the best scoring combination
• Rule Refinement: refine the acquired rules via a process of interaction with bilingual informants
• Morphology Learning• Word and Phrase bilingual lexicon acquisition
September 7, 2005 AVENUE Project 6
AVENUE Architecture
Learning Module
Transfer Rules
{PP,4894};;Score:0.0470PP::PP [NP POSTP] -> [PREP NP]((X2::Y1)(X1::Y2))
Translation Lexicon
Run Time Transfer System
Lattice Decoder
English Language Model
Word-to-Word Translation Probabilities
Word-aligned elicited data
September 7, 2005 AVENUE Project 7
Learning Transfer-Rules for Languages with Limited Resources
• Rationale:– Bilingual native informant(s) can translate and align a
small pre-designed elicitation corpus, using elicitation tool– Elicitation corpus designed to be typologically and
structurally comprehensive and compositional– Transfer-rule engine and new learning approach support
acquisition of generalized transfer-rules from the data
September 7, 2005 AVENUE Project 8
Transfer Rule Formalism
Type informationPart-of-speech/constituent
informationAlignments
x-side constraints
y-side constraints
xy-constraints, e.g. ((Y1 AGR) = (X1 AGR))
;SL: the old man, TL: ha-ish ha-zaqen
NP::NP [DET ADJ N] -> [DET N DET ADJ]((X1::Y1)(X1::Y3)(X2::Y4)(X3::Y2)
((X1 AGR) = *3-SING)((X1 DEF = *DEF)((X3 AGR) = *3-SING)((X3 COUNT) = +)
((Y1 DEF) = *DEF)((Y3 DEF) = *DEF)((Y2 AGR) = *3-SING)((Y2 GENDER) = (Y4 GENDER)))
September 7, 2005 AVENUE Project 9
Rule Learning - Overview
• Goal: Acquire Syntactic Transfer Rules• Use available knowledge from the source
side (grammatical structure)• Three steps:
1. Flat Seed Generation: first guesses at transfer rules; flat syntactic structure
2. Compositionality: use previously learned rules to add hierarchical structure
3. Constraint Learning: refine rules by learning appropriate feature constraints
September 7, 2005 AVENUE Project 10
Flat Seed Rule Generation
Learning Example: NP
Eng: the big apple
Heb: ha-tapuax ha-gadol
Generated Seed Rule:
NP::NP [ART ADJ N] [ART N ART ADJ]
((X1::Y1)
(X1::Y3)
(X2::Y4)
(X3::Y2))
September 7, 2005 AVENUE Project 11
CompositionalityInitial Flat Rules: S::S [ART ADJ N V ART N] [ART N ART ADJ V P ART N]
((X1::Y1) (X1::Y3) (X2::Y4) (X3::Y2) (X4::Y5) (X5::Y7) (X6::Y8))
NP::NP [ART ADJ N] [ART N ART ADJ]
((X1::Y1) (X1::Y3) (X2::Y4) (X3::Y2))
NP::NP [ART N] [ART N]
((X1::Y1) (X2::Y2))
Generated Compositional Rule:
S::S [NP V NP] [NP V P NP]
((X1::Y1) (X2::Y2) (X3::Y4))
September 7, 2005 AVENUE Project 12
Constraint LearningInput: Rules and their Example Sets
S::S [NP V NP] [NP V P NP] {ex1,ex12,ex17,ex26}
((X1::Y1) (X2::Y2) (X3::Y4))
NP::NP [ART ADJ N] [ART N ART ADJ] {ex2,ex3,ex13}
((X1::Y1) (X1::Y3) (X2::Y4) (X3::Y2))
NP::NP [ART N] [ART N] {ex4,ex5,ex6,ex8,ex10,ex11}
((X1::Y1) (X2::Y2))
Output: Rules with Feature Constraints:
S::S [NP V NP] [NP V P NP]
((X1::Y1) (X2::Y2) (X3::Y4)
(X1 NUM = X2 NUM)
(Y1 NUM = Y2 NUM)
(X1 NUM = Y1 NUM))
September 7, 2005 AVENUE Project 13
AVENUE Prototypes
• General XFER framework under development for past three years
• Prototype systems so far:– German-to-English, Spanish-to-English– Hindi-to-English, Hebrew-to-English
• In progress or planned:– Mapudungun-to-Spanish– Quechua-to-Spanish– Arabic-to-English– Native-Brazilian languages to Brazilian Portuguese
September 7, 2005 AVENUE Project 14
Morphology Learning
• Unsupervised learning of morphemes and their function from raw monolingual data– Segmentation of words into morphemes– Identification of morphological paradigms
(inflections and derivations)– Learning association between morphemes
and their function in the language
September 7, 2005 AVENUE Project 15
Ø.ed.ly11cleardirectpresentquiet…
Ø.ed.ing.ly6clearopenpresentTotal
ed.ly12bodiclearcorrectquiet…
Ø.ed.ing.ly.s4clearopen…
Ø.ed.ing201aidcleardefenddeliver…
d.ded.ding27aiboardefen…
d.ded.ding.ds19adboardefen…
Ø.ed.ing.s106cleardefendopenpresent…
Morphology LearningAVENUE Approach:
•Organize the raw data in the form of a network of paradigm candidate schemes•Search the network for a collection of schemes that represent true morphology paradigms of the language•Learn mappings between the schemes and features/functions using minimal pairs of elicited data•Construct analyzer based on the collection of schemes and the acquired function mappings
September 7, 2005 AVENUE Project 16
a.as.o.os43
africancasjurídicl...
a.as.i.o.os.sandra.tanier.ter.tro.trol
1cas
a.as.os50
afectadcasjurídicl...
a.as.o59cascitadjurídicl...
a.o.os105impuestindonesiitalianjurídic...
a.as199huelgincluidindustriinundad...
a.os134impedidimpuestindonesiinundad...
as.os68cas
implicadinundadjurídic...
a.o214idindi
indonesiinmediat...
as.o85internjurídicjustl...
a.tro2cascen
a1237huelgibidiglesi...
as404huelghuelguistincluidindustri...
os534
humorístichumanhígadimpedid...
o1139hubhughumanhuyend...
tro16catascecencua...
as.o.os54cas
implicadjurídicl...
Figure : Hierarchical scheme lattice automatically derived from a Spanish newswire corpus of 40,011 words and 6,975 unique types.
o.os268humanimplicadindici
indocumentad...
September 7, 2005 AVENUE Project 17
Automated Rule Refinement
• Rationale:– Bilingual informants can identify translation errors
and pinpoint the errors– A sophisticated trace of the translation path can
identify likely sources for the error and do “Blame Assignment”
– Rule Refinement operators can be developed to modify the underlying translation grammar (and lexicon) based on characteristics of the error source:
• Add or delete feature constraints from a rule• Bifurcate a rule into two rules (general and specific)• Add or correct lexical entries
September 7, 2005 AVENUE Project 18
New Research Directions• Automatic Transfer Rule Learning:
– In the “large-data” scenario: from large volumes of uncontrolled parallel text automatically word-aligned
– In the absence of morphology or POS annotated lexica
– Learning mappings for non-compositional structures– Effective models for rule scoring for
• Decoding: using scores at runtime• Pruning the large collections of learned rules
– Learning Unification Constraints – VSL
• Integrated Xfer Engine and Decoder– Improved models for scoring tree-to-tree mappings,
integration with LM and other knowledge sources in the course of the search
September 7, 2005 AVENUE Project 19
Missing Science
• Monolingual learning tasks:– Learning morphology: morphemes and their meaning– Learning syntactic and semantic structures:
grammar induction• Bilingual Learning Tasks:
– Automatic acquisition of word and phrase translation lexicons
– Learning structural mappings (syntactic, semantic, non-compositional)
• Models that effectively combine learned symbolic knowledge with statistical information: new “decoders”
September 7, 2005 AVENUE Project 20
September 7, 2005 AVENUE Project 21
English-Chinese Example
September 7, 2005 AVENUE Project 22
English-Hindi Example
September 7, 2005 AVENUE Project 23
Spanish-Mapudungun Example
September 7, 2005 AVENUE Project 24
English-Arabic Example
September 7, 2005 AVENUE Project 25
Transfer Rule Formalism (II)
Value constraints
Agreement constraints
;SL: the old man, TL: ha-ish ha-zaqen
NP::NP [DET ADJ N] -> [DET N DET ADJ]((X1::Y1)(X1::Y3)(X2::Y4)(X3::Y2)
((X1 AGR) = *3-SING)((X1 DEF = *DEF)((X3 AGR) = *3-SING)((X3 COUNT) = +)
((Y1 DEF) = *DEF)((Y3 DEF) = *DEF)((Y2 AGR) = *3-SING)((Y2 GENDER) = (Y4 GENDER)))
September 7, 2005 AVENUE Project 26
AVENUE PartnersLanguage Country Institutions
Mapudungun (in place)
Chile Universidad de la Frontera, Institute for Indigenous Studies, Ministry of Education
Quechua(discussion)
Peru Ministry of Education
Aymara(discussion)
Bolivia, Peru Ministry of Education
September 7, 2005 AVENUE Project 27
The Transfer EngineAnalysis
Source text is parsed into its grammatical structure. Determines transfer application ordering.
Example:
他 看 书。 (he read book)
S
NP VP
N V NP
他 看 书
TransferA target language tree is created by reordering, insertion, and deletion.
S
NP VP
N V NP
he read DET N
a book
Article “a” is inserted into object NP. Source words translated with transfer lexicon.
GenerationTarget language constraints are checked and final translation produced.
E.g. “reads” is chosen over “read” to agree with “he”.
Final translation:
“He reads a book”
September 7, 2005 AVENUE Project 28
Seeded VSL: Some Open Issues
• Three types of constraints:– X-side constrain applicability of rule– Y-side assist in generation– X-Y transfer features from SL to TL
• Which of the three types improves translation performance?– Use rules without features to populate lattice, decoder will select
the best translation…– Learn only X-Y constraints, based on list of universal projecting
features• Other notions of version-spaces of feature constraints:
– Current feature learning is specific to rules that have identical transfer components
– Important issue during transfer is to disambiguate among rules that have same SL side but different TL side – can we learn effective constraints for this?
September 7, 2005 AVENUE Project 29
Examples of Learned Rules (Hindi-to-English)
{NP,14244}
;;Score:0.0429
NP::NP [N] -> [DET N]
(
(X1::Y2)
)
{NP,14434}
;;Score:0.0040
NP::NP [ADJ CONJ ADJ N] ->
[ADJ CONJ ADJ N]
(
(X1::Y1) (X2::Y2)
(X3::Y3) (X4::Y4)
)
{PP,4894};;Score:0.0470PP::PP [NP POSTP] -> [PREP NP]((X2::Y1)(X1::Y2))
September 7, 2005 AVENUE Project 30
XFER MT for Hebrew-to-English• Two month intensive effort to apply our XFER approach
to the development of a Hebrew-to-English MT system• Challenges:
– No large parallel corpus– Limited coverage translation lexicon– Rich Morphology: incomplete analyzer available
• Accomplished:– Collected available resources, establish methodology for
processing Hebrew input– Translated and aligned Elicitation Corpus– Learned XFER rules– Developed (small) manual XFER grammar as a point of
comparison– System debugging and development– Evaluated performance on unseen test data using
automatic evaluation metrics
September 7, 2005 AVENUE Project 31
Transfer Engine
English Language Model
Transfer Rules{NP1,3}NP1::NP1 [NP1 "H" ADJ] -> [ADJ NP1]((X3::Y1) (X1::Y2) ((X1 def) = +) ((X1 status) =c absolute) ((X1 num) = (X3 num)) ((X1 gen) = (X3 gen)) (X0 = X1))
Translation Lexicon
N::N |: ["$WR"] -> ["BULL"]((X1::Y1) ((X0 NUM) = s) ((Y0 lex) = "BULL"))
N::N |: ["$WRH"] -> ["LINE"]((X1::Y1) ((X0 NUM) = s) ((Y0 lex) = "LINE"))
Hebrew Input
בשורה הבאה
Decoder
English Output
in the next line
Translation Output Lattice
(0 1 "IN" @PREP)(1 1 "THE" @DET)(2 2 "LINE" @N)(1 2 "THE LINE" @NP)(0 2 "IN LINE" @PP)(0 4 "IN THE NEXT LINE" @PP)
Preprocessing
Morphology
September 7, 2005 AVENUE Project 32
Morphology Example
• Input word: B$WRH
0 1 2 3 4 |--------B$WRH--------| |-----B-----|$WR|--H--| |--B--|-H--|--$WRH---|
September 7, 2005 AVENUE Project 33
Morphology ExampleY0: ((SPANSTART 0) Y1: ((SPANSTART 0) Y2: ((SPANSTART 1) (SPANEND 4) (SPANEND 2) (SPANEND 3) (LEX B$WRH) (LEX B) (LEX $WR) (POS N) (POS PREP)) (POS N) (GEN F) (GEN M) (NUM S) (NUM S) (STATUS ABSOLUTE)) (STATUS ABSOLUTE))
Y3: ((SPANSTART 3) Y4: ((SPANSTART 0) Y5: ((SPANSTART 1) (SPANEND 4) (SPANEND 1) (SPANEND 2) (LEX $LH) (LEX B) (LEX H) (POS POSS)) (POS PREP)) (POS DET))
Y6: ((SPANSTART 2) Y7: ((SPANSTART 0) (SPANEND 4) (SPANEND 4) (LEX $WRH) (LEX B$WRH) (POS N) (POS LEX)) (GEN F) (NUM S) (STATUS ABSOLUTE))
September 7, 2005 AVENUE Project 34
Sample Output (dev-data)
maxwell anurpung comes from ghana for israel four years ago and since worked in cleaning in hotels in eilat
a few weeks ago announced if management club hotel that for him to leave israel according to the government instructions and immigration police
in a letter in broken english which spread among the foreign workers thanks to them hotel for their hard work and announced that will purchase for hm flight tickets for their countries from their money
September 7, 2005 AVENUE Project 35
Evaluation Results
• Test set of 62 sentences from Haaretz newspaper, 2 reference translations
System BLEU NIST P R METEOR
No Gram 0.0616 3.4109 0.4090 0.4427 0.3298
Learned 0.0774 3.5451 0.4189 0.4488 0.3478
Manual 0.1026 3.7789 0.4334 0.4474 0.3617
September 7, 2005 AVENUE Project 36
Future Directions• Continued work on automatic rule learning (especially
Seeded Version Space Learning)– Use Hebrew and Hindi systems as test platforms for
experimenting with advanced learning research• Rule Refinement via interaction with bilingual speakers• Developing a well-founded model for assigning scores
(probabilities) to transfer rules• Redesigning and improving decoder to better fit the
specific characteristics of the XFER model• Improved leveraging from manual grammar resources• MEMT with improved
– Combination of output from different translation engines with different confidence scores
– strong decoding capabilities
September 7, 2005 AVENUE Project 37
Flat Seed Generation
Create a transfer rule that is specific to the sentence pair, but abstracted to the POS level. No syntactic structure.
Element Source
SL POS sequence f-structure
TL POS sequence TL dictionary, aligned SL words
Type information corpus, same on SL and TL
Alignments informant
x-side constraints f-structure
y-side constraints TL dictionary, aligned SL words (list of projecting features)
September 7, 2005 AVENUE Project 38
Compositionality - Overview
• Traverse the c-structure of the English sentence, add compositional structure for translatable chunks
• Adjust constituent sequences, alignments
• Remove unnecessary constraints, i.e. those that are contained in the lower-level rule
September 7, 2005 AVENUE Project 39
Seeded Version Space Learning: Overview
• Goal: add appropriate feature constraints to the acquired rules• Methodology:
– Preserve general structural transfer– Learn specific feature constraints from example set
• Seed rules are grouped into clusters of similar transfer structure (type, constituent sequences, alignments)
• Each cluster forms a version space: a partially ordered hypothesis space with a specific and a general boundary
• The seed rules in a group form the specific boundary of a version space
• The general boundary is the (implicit) transfer rule with the same type, constituent sequences, and alignments, but no feature constraints
September 7, 2005 AVENUE Project 40
Seeded Version Space Learning: Generalization
• The partial order of the version space:Definition: A transfer rule tr1 is strictly more general than another transfer rule tr2 if all f-structures that are satisfied by tr2 are also satisfied by tr1.
• Generalize rules by merging them:– Deletion of constraint– Raising two value constraints to an agreement
constraint, e.g. ((x1 num) = *pl), ((x3 num) = *pl) ((x1 num) = (x3 num))
September 7, 2005 AVENUE Project 41
Seeded Version Space Learning
NP v det n NP VP …1. Group seed rules into version spaces as above.2. Make use of partial order of rules in version space. Partial order is defined
via the f-structures satisfying the constraints.3. Generalize in the space by repeated merging of rules:
1. Deletion of constraint2. Moving value constraints to agreement constraints, e.g.
((x1 num) = *pl), ((x3 num) = *pl) ((x1 num) = (x3 num)
4. Check translation power of generalized rules against sentence pairs
September 7, 2005 AVENUE Project 42
Seeded Version Space Learning:The Search
• The Seeded Version Space algorithm itself is the repeated generalization of rules by merging
• A merge is successful if the set of sentences that can correctly be translated with the merged rule is a superset of the union of sets that can be translated with the unmerged rules, i.e. check power of rule
• Merge until no more successful merges
September 7, 2005 AVENUE Project 43
Conclusions• Transfer rules (both manual and learned) offer
significant contributions that can complement existing data-driven approaches– Also in medium and large data settings?
• Initial steps to development of a statistically grounded transfer-based MT system with:– Rules that are scored based on a well-founded
probability model – Strong and effective decoding that incorporates the
most advanced techniques used in SMT decoding
• Working from the “opposite” end of research on incorporating models of syntax into “standard” SMT systems [Knight et al]
• Our direction makes sense in the limited data scenario
September 7, 2005 AVENUE Project 44
AVENUE Architecture
User
Learning Module
ElicitationProcess
TransferRule
Learning
TransferRules
Run-Time Module
SLInput
SL Parser
TransferEngine
TLGenerator
TLOutputDecoder
MorphologyPre-proc
September 7, 2005 AVENUE Project 45
Learning Transfer-Rules for Languages with Limited Resources
• Rationale:– Large bilingual corpora not available– Bilingual native informant(s) can translate and align a
small pre-designed elicitation corpus, using elicitation tool– Elicitation corpus designed to be typologically
comprehensive and compositional– Transfer-rule engine and new learning approach support
acquisition of generalized transfer-rules from the data
September 7, 2005 AVENUE Project 46
The Elicitation Corpus
• Translated, aligned by bilingual informant• Corpus consists of linguistically diverse
constructions• Based on elicitation and documentation work
of field linguists (e.g. Comrie 1977, Bouquiaux 1992)
• Organized compositionally: elicit simple structures first, then use them as building blocks
• Goal: minimize size, maximize linguistic coverage
September 7, 2005 AVENUE Project 47
The Transfer EngineAnalysis
Source text is parsed into its grammatical structure. Determines transfer application ordering.
Example:
他 看 书。 (he read book)
S
NP VP
N V NP
他 看 书
TransferA target language tree is created by reordering, insertion, and deletion.
S
NP VP
N V NP
he read DET N
a book
Article “a” is inserted into object NP. Source words translated with transfer lexicon.
GenerationTarget language constraints are checked and final translation produced.
E.g. “reads” is chosen over “read” to agree with “he”.
Final translation:
“He reads a book”
September 7, 2005 AVENUE Project 48
Transfer Rule Formalism
Type informationPart-of-speech/constituent
informationAlignments
x-side constraints
y-side constraints
xy-constraints, e.g. ((Y1 AGR) = (X1 AGR))
;SL: the man, TL: der Mann
NP::NP [DET N] -> [DET N]((X1::Y1)(X2::Y2)
((X1 AGR) = *3-SING)((X1 DEF = *DEF)((X2 AGR) = *3-SING)((X2 COUNT) = +)
((Y1 AGR) = *3-SING)((Y1 DEF) = *DEF)((Y2 AGR) = *3-SING)((Y2 GENDER) = (Y1 GENDER)))
September 7, 2005 AVENUE Project 49
Transfer Rule Formalism (II)
Value constraints
Agreement constraints
;SL: the man, TL: der MannNP::NP [DET N] -> [DET N]((X1::Y1)(X2::Y2)
((X1 AGR) = *3-SING)((X1 DEF = *DEF)((X2 AGR) = *3-SING)((X2 COUNT) = +)
((Y1 AGR) = *3-SING)((Y1 DEF) = *DEF)((Y2 AGR) = *3-SING)((Y2 GENDER) = (Y1 GENDER)))
September 7, 2005 AVENUE Project 50
Rule Learning - Overview
• Goal: Acquire Syntactic Transfer Rules• Use available knowledge from the source
side (grammatical structure)• Three steps:
1. Flat Seed Generation: first guesses at transfer rules; flat syntactic structure
2. Compositionality: use previously learned rules to add hierarchical structure
3. Seeded Version Space Learning: refine rules by generalizing with validation (learn appropriate feature constraints)
September 7, 2005 AVENUE Project 51
Examples of Learned Rules (I){NP,14244}
;;Score:0.0429
NP::NP [N] -> [DET N]
(
(X1::Y2)
)
{NP,14434}
;;Score:0.0040
NP::NP [ADJ CONJ ADJ N] ->
[ADJ CONJ ADJ N]
(
(X1::Y1) (X2::Y2)
(X3::Y3) (X4::Y4)
)
{PP,4894};;Score:0.0470PP::PP [NP POSTP] -> [PREP NP]((X2::Y1)(X1::Y2))
September 7, 2005 AVENUE Project 52
A Limited Data Scenario for Hindi-to-English
• Put together a scenario with “miserly” data resources:– Elicited Data corpus: 17589 phrases– Cleaned portion (top 12%) of LDC dictionary: ~2725
Hindi words (23612 translation pairs)– Manually acquired resources during the SLE:
• 500 manual bigram translations• 72 manually written phrase transfer rules• 105 manually written postposition rules• 48 manually written time expression rules
• No additional parallel text!!
September 7, 2005 AVENUE Project 53
Manual Grammar Development
• Covers mostly NPs, PPs and VPs (verb complexes)
• ~70 grammar rules, covering basic and recursive NPs and PPs, verb complexes of main tenses in Hindi (developed in two weeks)
September 7, 2005 AVENUE Project 54
Manual Transfer Rules: Example;; PASSIVE OF SIMPLE PAST (NO AUX) WITH LIGHT VERB;; passive of 43 (7b){VP,28}VP::VP : [V V V] -> [Aux V]( (X1::Y2) ((x1 form) = root) ((x2 type) =c light) ((x2 form) = part) ((x2 aspect) = perf) ((x3 lexwx) = 'jAnA') ((x3 form) = part) ((x3 aspect) = perf) (x0 = x1) ((y1 lex) = be) ((y1 tense) = past) ((y1 agr num) = (x3 agr num)) ((y1 agr pers) = (x3 agr pers)) ((y2 form) = part))
September 7, 2005 AVENUE Project 55
Manual Transfer Rules: Example
; NP1 ke NP2 -> NP2 of NP1; Ex: jIvana ke eka aXyAya; life of (one) chapter ; ==> a chapter of life;{NP,12}NP::NP : [PP NP1] -> [NP1 PP]( (X1::Y2) (X2::Y1); ((x2 lexwx) = 'kA'))
{NP,13}NP::NP : [NP1] -> [NP1]( (X1::Y1))
{PP,12}PP::PP : [NP Postp] -> [Prep NP]( (X1::Y2) (X2::Y1))
NP
PP NP1
NP P Adj N
N1 ke eka aXyAya
N
jIvana
NP
NP1 PP
Adj N P NP
one chapter of N1
N
life
September 7, 2005 AVENUE Project 56
Adding a “Strong” Decoder
• XFER system produces a full lattice• Edges are scored using word-to-word
translation probabilities, trained from the limited bilingual data
• Decoder uses an English LM (70m words)• Decoder can also reorder words or phrases (up
to 4 positions ahead)• For XFER(strong) , ONLY edges from basic XFER
system are used!
September 7, 2005 AVENUE Project 57
Testing Conditions
• Tested on section of JHU provided data: 258 sentences with four reference translations– SMT system (stand-alone)– EBMT system (stand-alone)– XFER system (naïve decoding)– XFER system with “strong” decoder
• No grammar rules (baseline)• Manually developed grammar rules• Automatically learned grammar rules
– XFER+SMT with strong decoder (MEMT)
September 7, 2005 AVENUE Project 58
Results on JHU Test Set (very miserly training data)System BLEU M-BLEU NIST
EBMT 0.058 0.165 4.22
SMT 0.093 0.191 4.64
XFER (naïve) man grammar
0.055 0.177 4.46
XFER (strong)
no grammar0.109 0.224 5.29
XFER (strong) learned grammar
0.116 0.231 5.37
XFER (strong) man grammar
0.135 0.243 5.59
XFER+SMT 0.136 0.243 5.65
September 7, 2005 AVENUE Project 59
Effect of Reordering in the Decoder
NIST vs. Reordering
4.8
4.9
5
5.1
5.2
5.3
5.4
5.5
5.6
5.7
0 1 2 3 4
reordering window
NIS
T s
core no grammar
learned grammar
manual grammar
MEMT: SFXER+ SMT
September 7, 2005 AVENUE Project 60
Observations and Lessons (I)• XFER with strong decoder outperformed SMT even
without any grammar rules in the miserly data scenario– SMT Trained on elicited phrases that are very short– SMT has insufficient data to train more discriminative
translation probabilities– XFER takes advantage of Morphology
• Token coverage without morphology: 0.6989• Token coverage with morphology: 0.7892
• Manual grammar currently somewhat better than automatically learned grammar– Learned rules did not yet use version-space learning– Large room for improvement on learning rules – Importance of effective well-founded scoring of learned rules
September 7, 2005 AVENUE Project 61
Observations and Lessons (II)
• MEMT (XFER and SMT) based on strong decoder produced best results in the miserly scenario.
• Reordering within the decoder provided very significant score improvements– Much room for more sophisticated grammar rules– Strong decoder can carry some of the reordering
“burden”
September 7, 2005 AVENUE Project 62
Conclusions• Transfer rules (both manual and learned) offer
significant contributions that can complement existing data-driven approaches– Also in medium and large data settings?
• Initial steps to development of a statistically grounded transfer-based MT system with:– Rules that are scored based on a well-founded
probability model – Strong and effective decoding that incorporates the
most advanced techniques used in SMT decoding
• Working from the “opposite” end of research on incorporating models of syntax into “standard” SMT systems [Knight et al]
• Our direction makes sense in the limited data scenario
September 7, 2005 AVENUE Project 63
Future Directions• Continued work on automatic rule learning
(especially Seeded Version Space Learning)• Improved leveraging from manual grammar
resources, interaction with bilingual speakers• Developing a well-founded model for assigning
scores (probabilities) to transfer rules• Improving the strong decoder to better fit the
specific characteristics of the XFER model• MEMT with improved
– Combination of output from different translation engines with different scorings
– strong decoding capabilities
September 7, 2005 AVENUE Project 64
Rule Learning - Overview
• Goal: Acquire Syntactic Transfer Rules• Use available knowledge from the source
side (grammatical structure)• Three steps:
1. Flat Seed Generation: first guesses at transfer rules; no syntactic structure
2. Compositionality: use previously learned rules to add structure
3. Seeded Version Space Learning: refine rules by generalizing with validation
September 7, 2005 AVENUE Project 65
Flat Seed Generation
Create a transfer rule that is specific to the sentence pair, but abstracted to the POS level. No syntactic structure.
Element Source
SL POS sequence f-structure
TL POS sequence TL dictionary, aligned SL words
Type information corpus, same on SL and TL
Alignments informant
x-side constraints f-structure
y-side constraints TL dictionary, aligned SL words (list of projecting features)
September 7, 2005 AVENUE Project 66
Flat Seed Generation - Example
The highly qualified applicant did not accept the offer.Der äußerst qualifizierte Bewerber nahm das Angebot nicht an.
((1,1),(2,2),(3,3),(4,4),(6,8),(7,5),(7,9),(8,6),(9,7))
S::S [det adv adj n aux neg v det n] -> [det adv adj n v det n neg vpart](;;alignments:(x1:y1)(x2::y2)(x3::y3)(x4::y4)(x6::y8)(x7::y5)(x7::y9)(x8::y6)(x9::y7));;constraints:((x1 def) = *+) ((x4 agr) = *3-sing) ((x5 tense) = *past) …. ((y1 def) = *+) ((y3 case) = *nom) ((y4 agr) = *3-sing) …. )
September 7, 2005 AVENUE Project 67
Compositionality - Overview
• Traverse the c-structure of the English sentence, add compositional structure for translatable chunks
• Adjust constituent sequences, alignments• Remove unnecessary constraints, i.e. those
that are contained in the lower-level rule• Adjust constraints: use f-structure of correct
translation vs. f-structure of incorrect translations to introduce context constraints
September 7, 2005 AVENUE Project 68
Compositionality - Example
S::S [det adv adj n aux neg v det n] -> [det adv adj n v det n neg vpart](;;alignments:(x1:y1)(x2::y2)(x3::y3)(x4::y4)(x6::y8)(x7::y5)(x7::y9)(x8::y6)(x9::y7));;constraints:((x1 def) = *+) ((x4 agr) = *3-sing) ((x5 tense) = *past) …. ((y1 def) = *+) ((y3 case) = *nom) ((y4 agr) = *3-sing) …. )
S::S [NP aux neg v det n] -> [NP v det n neg vpart](;;alignments:(x1::y1)(x3::y5)(x4::y2)(x4::y6)(x5::y3)(x6::y4);;constraints:((x2 tense) = *past) …. ((y1 def) = *+) ((y1 case) = *nom) …. )
NP::NP [det AJDP n]-> [det ADJP n]
((x1::y1)…((y3 agr) = *3-sing)((x3 agr = *3-sing)
….)
September 7, 2005 AVENUE Project 69
Seeded Version Space Learning: Overview
• Goal: further generalize the acquired rules• Methodology:
– Preserve general structural transfer– Consider relaxing specific feature constraints
• Seed rules are grouped into clusters of similar transfer structure (type, constituent sequences, alignments)
• Each cluster forms a version space: a partially ordered hypothesis space with a specific and a general boundary
• The seed rules in a group form the specific boundary of a version space
• The general boundary is the (implicit) transfer rule with the same type, constituent sequences, and alignments, but no feature constraints
September 7, 2005 AVENUE Project 70
Seeded Version Space Learning
NP v det n NP VP …1. Group seed rules into version spaces as above.2. Make use of partial order of rules in version space. Partial order is defined
via the f-structures satisfying the constraints.3. Generalize in the space by repeated merging of rules:
1. Deletion of constraint2. Moving value constraints to agreement constraints, e.g.
((x1 num) = *pl), ((x3 num) = *pl) ((x1 num) = (x3 num)
4. Check translation power of generalized rules against sentence pairs
September 7, 2005 AVENUE Project 71
Seeded Version Space Learning: Example
S::S [NP aux neg v det n] -> [NP v det n neg vpart](;;alignments:(x1::y1)(x3::y5)(x4::y2)(x4::y6)(x5::y3)(x6::y4);;constraints:((x2 tense) = *past) …. ((y1 def) = *+) ((y1 case) = *nom) ((y1 agr) = *3-sing) … )((y3 agr) = *3-sing) ((y4 agr) = *3-sing)… )
S::S [NP aux neg v det n] -> [NP v det n neg vpart](;;alignments:(x1::y1)(x3::y5)(x4::y2)(x4::y6)(x5::y3)(x6::y4);;constraints:((x2 tense) = *past) …((y1 def) = *+) ((y1 case) = *nom) ((y1 agr) = *3-plu) …((y3 agr) = *3-plu) ((y4 agr) = *3-plu)… )
S::S[NP aux neg v det n] -> [NP n det n neg vpart](;;alignments:(x1::y1)(x3::y5)(x4::y2)(x4::y6)(x5::y3)(x6::y4);;constraints:((x2 tense) = *past) …((y1 def) = *+) ((y1 case) = *nom)((y4 agr) = (y3 agr))… )
September 7, 2005 AVENUE Project 72
Preliminary Evaluation
• English to German• Corpus of 141 ADJPs, simple NPs and
sentences• 10-fold cross-validation experiment• Goals:
– Do we learn useful transfer rules?– Does Compositionality improve
generalization?– Does VS-learning improve generalization?
September 7, 2005 AVENUE Project 73
Summary of Results
• Average translation accuracy on cross-validation test set was 62%
• Without VS-learning: 43%• Without Compositionality: 57%• Average number of VSs: 24• Average number of sents per VS: 3.8• Average number of merges per VS: 1.6• Percent of compositional rules: 34%
September 7, 2005 AVENUE Project 74
Conclusions
• New paradigm for learning transfer rules from pre-designed elicitation corpus
• Geared toward languages with very limited resources
• Preliminary experiments validate approach: compositionality and VS-learning improve generalization
September 7, 2005 AVENUE Project 75
Future Work
1. Larger, more diverse elicitation corpus2. Additional languages (Mapudungun…)3. Less information on TL side4. Reverse translation direction5. Refine the various algorithms:
• Operators for VS generalization• Generalization VS search• Layers for compositionality
6. User interactive verification
September 7, 2005 AVENUE Project 76
Seeded Version Space Learning: Generalization
• The partial order of the version space:Definition: A transfer rule tr1 is strictly more general than another transfer rule tr2 if all f-structures that are satisfied by tr2 are also satisfied by tr1.
• Generalize rules by merging them:– Deletion of constraint– Raising two value constraints to an agreement
constraint, e.g. ((x1 num) = *pl), ((x3 num) = *pl) ((x1 num) = (x3 num))
September 7, 2005 AVENUE Project 77
Seeded Version Space Learning: Merging Two Rules
Merging algorithm proceeds in three steps. To merge tr1 and tr2 into trmerged:
1. Copy all constraints that are both in tr1 and tr2 into trmerged
2. Consider tr1 and tr2 separately. For the remaining constraints in tr1 and tr2 , perform all possible instances of raising value constraints to agreement constraints.
3. Repeat step 1.
September 7, 2005 AVENUE Project 78
Seeded Version Space Learning:The Search
• The Seeded Version Space algorithm itself is the repeated generalization of rules by merging
• A merge is successful if the set of sentences that can correctly be translated with the merged rule is a superset of the union of sets that can be translated with the unmerged rules, i.e. check power of rule
• Merge until no more successful merges