Title: Coupling between ASR and MT in SpeechtoSpeech Translation
1Coupling between ASR and MT in Speech-to-Speech
Translation
- Arthur Chan
- Prepared for
- Advanced Machine Translation Seminar
2This Seminar (35 pages)
- Introduction (6 slides)
- Ringgers categorization of Coupling between ASR
and NLU (7 slides) - Interfaces in Loose Coupling
- 1 best and N-best (5 slides)
- Lattices/Confusion Network/Confidence Estimation
(9 slides) - Results from literature (4 slides)
- Tight Coupling
- Neys Theory and 2 methods of Implementation (4
slides) - (? Sorry, no FST approaches will be discussed)
- Many Bonus Material at the back
3History of this presentation
- V1
- Draft finished in Mar 1st
- Tanjas comment
- Direct modeling could be skipped.
- We could focus on telling why/ASR
- Generates the current outputs
- Issues in MT searching could be ignored.
4History of this presentation (cont.)
- V2 V4
- Followed Tanjas comment and finished in Mar 19th
. - Reviewers comment
- Too long (70 pages)
- Neys search formulation is too difficult to
follow - V5 V6
- Significantly trimmed down the presentation
- Moved a lot of things to the backup section.
- V7
- Incorporated some comments from Alon, Stephan and
the class.
54 papers on Coupling of Speech-to-Speech
Translation
- H. Ney, Speech translation Coupling of
recognition and translation, in Proc. ICASSP,
1999. - S.Saleem, S. C. Jou, S. Vogel, and T. Schultz,
Using word lattice information for a tighter
coupling in speech translation systems, in Proc.
ICSLP, 2004. - V.H. Quan et al., Integrated N-best re-ranking
for spoken language translation, in In
EuroSpeech, 2005. - N. Bertoldi and M. Federico, A new decoder for
spoken language translation based on confusion
networks, in IEEE ASRU Workshop, 2005.
6A Conceptual Model of Speech-to-Speech Translation
Speech Recognizer
Machine Translator
Speech Synthesizer
Decoding Result(s)
Translation
waveforms
waveforms
7Motivation of Tight Coupling between ASR and MT
- One best of ASR could be wrong
- MT could be benefited from wide range of
supplementary information provided by ASR - N-best list
- Lattice
- Sentenced/Word-based Confidence Scores
- E.g. Word posterior probability
- Confusion network
- Or consensus decoding (Mangu 1999)
- MT quality may depend on WER of ASR (?)
8Scope of this talk.
Speech Recognizer
Machine Translator
Speech Synthesizer
1-best?
Translation
N-best?
waveforms
waveforms
Lattice?
Confusion network?
Loose Coupling/ Tight Coupling
9Topics Covered Today
- The concept of Coupling
- Tightness of coupling between ASR and
Technology X. (Ringger 95) - Two questions
- What could ASR provide in loose coupling?
- Discussion of interfaces between ASR and MT in
loose coupling - What is the status of tight coupling?
- Neys Formulation
10Topics not covered
- Direct Modeling
- Use both features in ASR and MT
- Some referred as ASR and MT unification
- FST approaches
- V7 I only read two papers and couldnt do the
justcice. - Implication of the MT search algorithms on the
coupling - Generation of speech from text.
11The Concept of Coupling
12Classification of Coupling of ASR and Natural
Language Understanding (NLU)
- Proposed in Ringger 95, Harper 94
- 3 Dimensions of ASR/NLU
- Complexity of the search algorithm
- Simple N-gram?
- Incrementality of the coupling
- On-line? Left-to-right?
- Tightness of the coupling
- Tight? Loose? Semi-tight?
13Tightness of Coupling
Tight
Semi-Tight
Loose
14Notes
- Semi-tight coupling could appear as
- Feedback loop between ASR and Technology X for
the whole utterance of speech - Or Feedback loop between ASR and Technology X for
every frame. - The Ringger framework
- A good way to understand how speech-based system
is developed
15Example 1 LM
- Someone asserts that ASR has to be used with
13-grams. - In tight-coupling,
- A search will be devised to search for the best
word sequence with best acoustic score 13 gram
likelihood - In loose coupling
- A simple search will be used to generate some
outputs (N-best list, lattice etc.), - 13-gram will then use to rescore the output.
- In semi-tight coupling
- 1, A simple search will be used to generate
results - 2, 13 gram will be applied at the word-end only
(but exact history will not be stored)
16Example 2 Higher order AM
- Segmental model assume obs. probability is not
conditionally independent. - Someone assert that segmental model is better
than just HMM. - Tight coupling Direct search of the best word
sequence using segmental model. - Loose coupling Use segmental model to rescore
- Semi-tight coupling Hybrid HMM-Segmental model
algorithm?
17Summary of Coupling between ASR and NLU
18Implication on ASR/MT coupling
- Generalize many systems
- Loose coupling
- Any system which uses 1-best, n-best, lattice, or
other inputs for 1-way module communication - (Bertoldi 2005)
- CMU System (Saleem 2004)
- Tight coupling
- (Ney 1999)
- Semi-tight coupling
- (Quan 2005)
19Interfaces in Loose Coupling1-best and N-best
20Perspectives
- ASR outputs
- 1-best results
- N-best results
- Lattice
- Consensus network.
- Confidence scores
- How ASR generate these outputs?
- Why they are generated?
- What if there are multiple ASRs?
- (and what if their results are combined?)
- Note we are talking about state-lattice now,
not word-lattice. ?
21Origin of the 1-best.
- Decoding of HMM-based ASR
- Searching the best path in a huge HMM-state
lattice. - 1-best ASR result
- The best path one could find from backtracking.
- State Lattice in ASR (Next page)
22(No Transcript)
23Note on 1-best in ASR
- Most of the time 1-best Word Sequence
- Why?
- In LVCSR, storing the backtracking pointer table
for state sequence takes a lot of memory (even
nowadays) - Compare this with the number of frames of score
one need to be stored - Usually a backtrack pointer storing
- The previous words before the current word
- Clever structure dynamically allocate
back-tracking pointer table.
24What is N-best list?
- Traceback not only from the 1st -best, also from
the 2nd best and 3rd best, etc. - Pathway
- Directly from search backtrack pointer table
- Exact N-best algorithm (Chow 90)
- Word pair N-best algorithm (Chow 91)
- A search using Viterbi score as heuristic (Chow
92) - Generate lattice first, then generate N-best from
lattice
25Interfaces in Loose CouplingLattice, Consensus
Network and Confidence Estimation
26What is Lattice?
- A word-based lattice
- A compact representation of state-lattice
- Only word node (or link) are involved
- Difference between N-best and Lattice
- Lattice could be compact representation of N-best
list.
27(No Transcript)
28How lattice is generated?
- From the decoding backtracking pointer table
- Only record all the links between word nodes.
- From N-best list
- Become a compact representation of N-best
- sometimes spurious link will be introduced
- Some complicated issue
- Triphone contexts
- Cause a lot of complicated issue
- When lattice is too large
- You want to trim it.
29Conclusions on lattices
- Lattice generation itself could be a complicated
issue - Sometimes, what post-processing stage (e.g. MT)
will get is pre-filtered, pre-processed results.
30Confusion Network and Consensus Hypothesis
- Confusion Network
- Or Sausage Network.
- Or Consensus Network
31Special Properties
- More local than lattice
- One can apply simple criteria to find the best
results - E.g. consensus decoding is to apply
word-posterior probability on confusion network. - More tractable
- In terms of size
32Note on Consensus Network
- Note
- Time information might not be preserved in
confusion network - The similarity function directly affect the final
output of the consensus network. - Other ways to generate confusion network
- From the N-best list
- Using Rover.
- A mixture of voting and adding confidence of word
33Confidence Measure
- Anything other than likelihood which could tell
whether the answer is useful - E.g.
- Word posterior probability
- P(WA)
- Usually compute using lattices
- Language model backoff mode
- Other posterior probabilities (frame, sentence)
34Interfaces in Loose CouplingResults from the
Literature
35General Note
- Coupling in SST is still pretty new
- Papers are chosen according to whether some
outputs have been used - Other techniques such as direct modeling might be
mixed into the papers.
36N-best list (Quan 2005)
- Using N-best list for reranking
- Interpolation weights of AM and TM are then
optimized. - Summary
- Reranking gives improvements.
37Lattices CMU results (Saleem 2004)
- Summary of results
- Lattice word error rate improved when lattice
density improves - Lattice density and Weight on Acoustic scores
turns out to be an important parameter to tune - Too large and small could hurt.
38Consensus Network
- Bertoldi 2005 is probably the only work on
confusion-network based method - Summary of results
- When direct modeling is applied
- Consensus Network doesnt beat N-best method.
- Author argues for speed and simplicity of the
algorithm
39Confidence Does it help?
- According to Zhang 2006, Yes.
- Confidence Measure (CM) filtering is used to
filter out unnecessary results in N-best - Note The approaches used is quite different.
40Conclusion on Loose Coupling
- SR could give a rich set of outputs.
- It seems that it is still an unknown what type of
output should be used in pipeline. - Currently, it seem to lack of comprehensive
experimental studies on which method is the best.
- Usage of confusion network and confidence
estimation seem to be under-explored.
41Comments about Consensus Network
- From Stephan
- Reasons not using consensus networks now
- 1, the consensus network might occasionally give
spurious links in each sausage segment. - 2, lattices from the ASR teams could change from
time to time. MT teams need time to consume
them. - From Alon, Ralf and Stephan
- There are not much big reasons not to use
consensus network because essentially it is just
another type of network.
42Tight Coupling Theory and Practice
43Theory (Ney 1999)
Bayes Rule
Introduce f as hidden var.
Bayes Rule
Assume x doesnt depend on target lang.
Sum to Max
44Layman point of view
- Three factors
- Pr(e) target language model
- Pr(fe) translation model
- Pr(xf) acoustic model
- Note assumption has been made only the best
matching f for e is used.
45Comparison with SR
- In SR
- Pr(f) Source language model
- In Tight coupling
- Pr(fe), Pr(e) Translation model and Target
language model
46Algorithmic Point of View
- Brute Force Method Instead of incorporating LM
into standard Viterbi algorithm - Incoporating P(e) and P(fe)
- gt Very complicated
- The backup slides in the presentation has detail
about Neys implementations.
47Experimental Results in Matusov, Kanthak and Ney
2005
- Summary of the results
- Translation quality is only improved by tight
coupling when the lattice density is not high. - Same as Saleem 2004, incorporation of acoustic
scores help.
48Conclusion Possible Issues of tight coupling
- Possibilities
- In SR, source n-gram LM is very closed to the
best configuration. - The complexity of the algorithm is too high,
approximation is still necessary to make it work. - When the criterion in tight coupling is used. It
is possible that the LM and the TM need to be
jointly estimated. - The current approaches still havent really
implement tight-coupling - There might be bugs in the programs.
49Conclusion
- Two major issues in coupling of SST is discussed
- In loose coupling
- Consensus network and Confidence scoring is still
not fully utilized - In tight coupling
- The approach seem to be haunted by very high
complexity of search algorithm construction
50Discussion
- Ian It could be quite difficult to characterize
a relationship of WER and BLEU. - Alan ask Why not jointly optimize translation
model and acoustic model? - Arthur direct modeling could be useful
- Stephan (rephrase) will it really help?
51The End. Thanks.
52Literature
- 2006 Ruiqiang Zhang, Genichiro Kikui. Integration
of Speech Recognition and Machine Translation
Speech Recognition Word Lattice Translation.
Speech Communication. Vol.48, Issues 3-4 - H. Ney, Speech translation Coupling of
recognition and translation, in Proc. ICASSP,
1999. - E. Matusov, S.Kanthak, and H. Ney, On the
integration of speech recognition and statistical
machine translation, in Proc. InterSpeech, 2005. - S.Saleem, S. C. Jou, S. Vogel, and T. Schultz,
Using word lattice information for a tighter
coupling in speech translation systems, in Proc.
ICSLP, 2004. - V.H. Quan et al., Integrated N-best re-ranking
for spoken language translation, in In
EuroSpeech, 2005. - N. Bertoldi and M. Federico, A new decoder for
spoken language translation based on confusion
networks, in IEEE ASRU Workshop, 2005. - L. Mangu, E. Brill, A. Stolcke, Finding
consensus in speech recognition word error
minimization and other applications of confusion
networks, Computer Speech and Language 14(4),
373-400., (2000) - E. Ringger, A Robust Loose Coupling for Speech
Recognition and Natural Language Understanding,
1995
53Backup Slides
54Saleems results
55LWER against Lattice Density
56Modified Bleu scores against lattice density
57Optimal density and score weight based on
Utterance Length.
58Some Lattice-specific Issue
59How lattice is generated when there are phone
contexts at the word end?
- Very complicated when phonetic context is
involved - Not only word-end needs to be stored but also the
phone contexts. - Lattice has the word identity as well as contexts
- Lattice can become very large.
60How this is resolved?
- Some used only approximate triphone to generate
lattice in first stage (BBN) - Some generate lattice even with full CD-phones
but convert it back to no-context lattices (RWTH) - Use the lattice with full CD phone contexts
(RWTH)
61What ASR folks do when lattice is still too large?
- Use some criteria to prune the lattice.
- Example Criteria
- Word posterior probability
- Application of another LM or AM, then filtering.
- General confidence score
- Maximum lattice density
- (number of words in lattice/number of words)
- Or generate an even more compact representation
than lattices - E.g. consensus network.
62Ney 99s Formulation of SSTs Search.
63Assumptions in Modeling
- Alignment Models (HMM)
- Acoustic Modeling
- Speech Recognizer will produce a word graph.
- Each link with word hypothesis covers the portion
of acoustic scores. (notation is confusing in
paper)
64Lexicon Modeling
- Further assumption from standard IBM models
- Target word is assumed to be dependent on
previous word - So, in fact, source LM is actually there.
65First Implementation Local Average Assumptions
- Local Average Assumptions
- P(xe) is used to capture the local
characteristic of the acoustic.
66Justification of Using Average Local Assumption
- Rephrased from Author (p.3 para 2)
- Lexicon modeling and language modeling will cause
f_j-1, f_j, f_j1 appear in the math. - In another words
- It is too complicated to carry out
- Computation advantage the local score could be
obtained just from the word graph but before
translation - gt Full translation strategy could still be
carried out
67Computation of P(xe)
- Make use of best source sequence
- Also refer to Wessel 98,
- A commonly used word posterior probability
algorithm for lattice - A forward-backward like procedure is used
68Second Method Monotone Alignment Assumption -
Network
69Monotone Alignment Assumption Formula for Text
Input
- Close-formed solution exist form DP O(JE2)
70Monotone Alignment Assumption Formula for
Speech Input
71How to make Monotone Assumptions work?
- Words needs to be reordered
- As part of search strategy.
- Does acoustic model assumption used?
- i.e. Are we talking about word lattice or still
state lattice? - Dont know, seems like we are actually talking
about word lattice. - Supported by Matusov 2005