Wednesday, September 7, 2011

Training Machine Translation with a Second-Order Taylor Approximation of Weighted Translation Instances

Title: Training Machine Translation with a Second-Order Taylor Approximation of Weighted Translation Instances
Speaker: Aaron Phillips
When: Tuesday, September 13, 12:00 Noon to 1:00pm.
Where: GHC 6501

Abstract: The Cunei Machine Translation Platform is an open-source MT system designed to model instances of translation. One of the challenges to this approach is effective training. We describe two techniques that improve the training procedure and allow us to leverage the strengths of instance-based modeling. First, during training we approximate our model with a second-order Taylor series. Second, we discount models based on the magnitude of their approximation. By reducing error in training, our model now consistently outperforms the standard SMT model with gains ranging from 0.51 to 3.77 BLEU on German-English and Czech-English test sets.

Monday, May 16, 2011

Syntax-to-Morphology Mapping in Factored Phrase-Based SMT (English and Turkish)

Title: Syntax-to-Morphology Mapping in Factored Phrase-Based
Statistical Machine Translation between English and Turkish

Speaker: Reyyan Yeniterzi

When: Tuesday, May 17 at 12:15pm
Where: GHC 6501

Abstract:

Motivated by the observation that many local and some nonlocal
syntactic structures in English essentially map to morphologically
complex words in Turkish, a new approach which is called
syntax-to-morphology mapping was introduced recently (Yeniterzi and
Oflazer, 2010). This approach maps syntactic structures in English to
complex words in Turkish directly. It mainly recognizes certain local
and nonlocal syntactic structures on the English side and packages
those structures and attach to heads to obtain parallel morphological
structures.

With the help of this method, one can identify and reorganize phrases
on the English side, to align English syntax to Turkish morphology.
Furthermore with this method, continuous and discontinuous variants of
certain (syntactic) source phrases can be conflated during the SMT
phrase extraction process. Since most function words encoding syntax
are now abstracted into complex tags, the length of the English
sentences can be dramatically reduced.

The initial experiments were performed on English-to-Turkish SMT
system. In this project, we built upon this initial system by doing
lexical reordering and data augmentation. Furthermore we also applied
syntax-to-morphology mapping to a Turkish-to-English SMT system for
the first time.

This is joint work with Kemal Oflazer from Qatar CMU. It was presented
in the Machine Translation and Morphologically-rich Languages Research
Workshop at Haifa, Israel in January, 2011.

Tuesday, May 3, 2011

Better Hypothesis Testing for Statistical Machine Translation: Controlling for Optimizer Instability

Title: Better Hypothesis Testing for Statistical Machine Translation:
Controlling for Optimizer Instability
Speaker: Jonathan Clark
When: Tuesday, 4/19 at Noon

Abstract:
In statistical machine translation, a researcher seeks to determine
whether some innovation (e.g., a new feature, model, or inference
algorithm) improves translation quality in comparison to a baseline
system. To answer this question, he runs an experiment to evaluate the
behavior of the two systems on held-out data. In this paper, we
consider how to make such experiments more statistically reliable. We
provide a systematic analysis of the effects of optimizer instability
(an extraneous variable that is seldom controlled for) on experimental
outcomes, and make recommendations for reporting results more
accurately.

This is joint work with Chris Dyer, Alon Lavie, and Noah Smith. It was
recently accepted for publication as an ACL short paper.

Wednesday, March 2, 2011

Qin Gao: Expanding parallel corpora for machine translation

Speaker: Qin Gao
When: at noon, March 8, 2011
Where: GHC 4405

We present an approach of expanding parallel corpora for machine translation. By utilizing Semantic role labeling (SRL) on one side of the language pair, we extract SRL substitution rules from existing parallel corpus. The rules are then used for generating new sentence pairs. An SVM classifier is built to filter the generated sentence pairs. The filtered corpus is used for training phrase-based translation models, which can be used directly in translation tasks or combined with baseline models. Experiment results on Chinese-English machine translation tasks show an average improvement of 0.45 BLEU and 1.22 TER points across 5 different NIST test sets.

Thursday, January 13, 2011

Machine Translation and Computer-Assisted Translation

Title: Prospects for Integrating Machine Translation and Computer-Assisted Translation in the Translation Industry

Speaker: Gregory M. Shreve from the Department of Modern and Classical Language Studies at Kent State University and colleagues
Location: GHC 6115
Time: 12:30 pm, 14 Jan 2011


The speaker's CV can be found at http://www.kent.edu/mcls/faculty/mcls_shreve.cfm.

Monday, December 13, 2010

Efficient Language Model Inference - Kenneth Heafield

Title: Efficient Language Model Inference
Who? Kenneth Heafield
When? Tuesday, December 21 @ Noon
Where? GHC 4405


In GHC 4405 at noon on Tuesday Dec 21, I will give a speaking
requirement talk on Efficient Language Model Inference. As this is also
a MT Lunch, there will be free lunch.

If you're using SRILM, come to my talk to reduce your memory consumption
by 86% while reducing CPU time by 16%. Users of IRSTLM should come for
the same reason; the code uses 42% less memory and 19% less CPU.

Language models are an important feature in speech, translation,
generation, IR, and other technologies. More training data and less
pruning generally lead to higher quality, but RAM is a limiting factor.
Further, systems consult language models so frequently that lookups
dominate CPU time.

This talk presents language modeling code with several optimizations to
improve time and space performance. Storing backoff information in
feature state reduces redundant lookups. Constructing known
distributions and biasing binary search speeds search and reduces page
faults. Memory mapping reduces load time. Bit level packing increases
locality. Stronger filtering removes n-grams that cannot be assembled
during decoding due to phrase and sentence constraints. The code is
currently integrated into Moses and being integrated into cdec and
Joshua. I will cover how my code works and how to use it in other
decoders.

Wednesday, October 6, 2010

Choosing the Right Evaluation for Machine Translation

Time: Noon on Tuesday, October 12
Place: GHC 6501 (usual location)

Title: Choosing the Right Evaluation for Machine Translation: an Examination of Annotator and Automatic Metric Performance on Human Judgment Tasks
Authors: Michael Denkowski and Alon Lavie

Abstract:
This work examines the motivation, design, and practical results of several types of human evaluation tasks for machine translation. In addition to considering annotator performance and task informativeness over multiple evaluations, we explore the practicality of tuning automatic evaluation metrics to each judgment type in a comprehensive experiment using the METEOR metric. We present results showing clear advantages of tuning to certain types of judgments and discuss causes of inconsistency when tuning to various judgment data, as well as sources of difficulty in the human evaluation tasks themselves.

This work will be presented at AMTA 2010.

Monday, September 13, 2010

Models for Synchronous Grammar Induction for Statistical Machine Translation

Title: Models for Synchronous Grammar Induction for Statistical Machine Translation
Presenters: Chris Dyer, LTI& Desai Chen, CSD undergraduate
Tuesday, September 14, at Noon to 1:30pm in GHC 6501.

Abstract: The last decade of research in Statistical Machine Translation (SMT) has seen rapid progress. The most successful methods have been based on synchronous context free grammars (SCFGs), which encode translational equivalences and license reordering between tokens in the source and target languages. Yet, while closely related language pairs can be translated with a high degree of precision now, the result for distant pairs is far from acceptable. In theory, however, the "right"' SCFG is capable of handling most, if not all, structurally divergent language pairs. This talk will report on the results of the 2010 Language Engineering Workshop held at Johns Hopkins University that the goal to focus on the crucial practical aspects of acquiring such SCFGs from bilingual, but otherwise unannotated, text. We started with existing algorithms for inducing unlabeled SCFGs (e.g. the popular Hiero model) and then used unsupervised learning methods to refine the syntactic constituents used in the translation rules of the grammar.

Monday, June 28, 2010

Syntax-to-Morphology Mapping in Factored Phrase-Based Statistical Machine Translation from English to Turkish

Tuesday, June 29 at Noon, in GHC 6501

Title: Syntax-to-Morphology Mapping in Factored Phrase-Based Statistical
Machine Translation from English to Turkish

Authors: Reyyan Yeniterzi and Kemal Oflazer


Abstract:

We present a novel scheme to apply factored phrase-based SMT to a
language pair with very disparate morphological structures. Our
approach relies on syntactic analysis on the source side (English) and
then encodes a wide variety of local and non-local syntactic
structures as complex structural tags which appear as additional
factors in the training data. On the target side (Turkish), we only
perform morphological analysis and disambiguation but treat the
complete complex morphological tag as a factor, instead of separating
morphemes. We incrementally explore capturing various syntactic
substructures as complex tags on the English side, and evaluate how
our translations improve in BLEU scores. Our maximal set of source and
target side transformations, coupled with some additional techniques,
provide an 39\% relative improvement from a baseline 17.08 to 23.78
BLEU, all averaged over 10 training and test sets. Now that the
syntactic analysis on the English side is available, we also
experiment with more long distance constituent reordering to bring the
English constituent order close to Turkish, but find that these
transformations do not provide any additional consistent tangible
gains when averaged over the 10 sets.

Wednesday, May 12, 2010

Chunk-Based EBMT

When: noon on May 18
Where: GHC 6501

Speaker: Jaedy Kim

Topic: Chunk-Based EBMT

Abstract: Corpus driven machine translation approaches such as Phrase-Based Statistical Machine Translation and Example-Based Machine Translation have been successful by using word alignment to find translation fragments for matched source parts in a bilingual training corpus.
However, they still cannot properly deal with systematic translation for insertion or deletion words between two distant languages.
In this work, we used syntactic chunks as translation units to alleviate this problem, improve alignments and show improvement in BLEU for Korean to English and Chinese to English translation tasks.

Monday, April 12, 2010

Generalized templates for EBMT

Speaker: Rashmi Gangadharaiah
Location: GHC 6501

Topic: Generalized templates for EBMT

Abstract:
-----------
Example-Based Machine Translation (EBMT), like other corpus based methods, requires substantial parallel training data. One way to reduce data requirements and improve translation quality is to generalize parts of the parallel corpus into translation templates. This automated generalization process requires clustering. In most clustering approaches the optimal number of clusters (N) is found empirically on a development set which often takes several days. We introduce a spectral clustering framework that automatically estimates the optimal N and removes unstable oscillating points. The new framework produces significant improvements in low-resource EBMT settings for English-to-French (~1.4 BLEU points), English-to-Chinese (~1 BLEU point), and English-to-Haitian (~2 BLEU points). The translation quality with templates created using automatically and empirically found best N were almost the same. By discarding “incoherent” points, a further boost in translation scores is observed, even above the empirically found N.

Monday, March 15, 2010

Two talks

(1) Greg Hanneman:
Title: The Stat-XFER Group Submission for WMT '10

Abstract:
Each year, the Workshop in Statistical Machine Translation collects state-of-the-art MT results for a variety of European language pairs via a shared translation task. In this talk, I will describe the CMU's Stat-XFER MT group submission to this year's WMT French--English track, our third submission to the WMT series, using the Joshua decoder. A large focus will be on new modeling decisions or system-building techniques that have changed from eariler submissions based on new research carried out in our group. I will also present some open questions facing builders of large-scale hierarchcial MT systems in general.


(2) Vamshi Ambati:
Title: Making sense of Crowd data for Machine Translation

Abstract:
Quality of crowd data is a common concern in crowd-sourcing approaches to data collection. When working with crowd data, the objectives are two-fold - maximizing the quality of data from non-experts, and minimizing the cost of annotation by pruning noisy annotators.
I will discuss our recent experiments in Machine Translation for selection of high quality crowd translations by explicitly modeling annotator reliability based on agreement with other submissions. I will also present some preliminary results in cost minimization and report their adaptation and feasibility to machine translation.

Thursday, February 18, 2010

Nonparametric Word Segmentation for Machine Translation

Speaker: Thuylinh Nguyen
Title: Nonparametric Word Segmentation for Machine Translation
Thursday 18 Feb 2010. 12-1:30pm in GHC 4405.

In this talk we present an unsupervised word segmentation for machine
translation. The model utilizes existing nonparametric monolingual
segmentations. The monolingual segmentation model and the bilingual word
alignment model are coupled so that source text segmentation optimizes
the one-to-one mapping with the target text. Often, there are words in
the source language that do not appear in target language and vise
versa. Our model therefore models source language word deletion and word
insertion. The experiments show improvements on Arabic-English and
Chinese-English translation tasks.

Wednesday, January 13, 2010

LoonyBin: Making Empirical MT Reproducible, Efficient, and Less Annoying

Speaker: Jonathan Clark
When: Tuesday, January 19 at Noon
Where: GHC 6501
What: Free Knowledge and Free Food
Title: LoonyBin: Making Empirical MT Reproducible, Efficient, and
Less Annoying

Abstract: Construction of machine translation systems has evolved into
a multi-stage workflow involving many complicated dependencies. Many
decoder distributions have addressed this by including monolithic
training scripts – train-factored-model.pl for Moses and mr_runmer.pl
for SAMT. However, such scripts can be tricky to modify for novel
experiments and typically have limited support for the variety of job
schedulers found on academic and commercial computer clusters. Further
complicating these systems are hyperparameters, which often cannot be
directly optimized by conventional methods requiring users to
determine which combination of values is best via trial and error. The
recently-released LoonyBin open-source workflow management tool
addresses these issues by providing: 1) a visual interface for the
user to create and modify workflows; 2) a well-defined logging
mechanism; 3) a script generator that compiles visual workflows into
shell scripts, and 4) the concept of Hyperworkflows, which intuitively
and succinctly encodes small experimental variations within a larger
workflow. We also describe the Machine Translation Toolpack for
LoonyBin, which exposes state-of-the-art machine translation tools as
drag-and-drop components within LoonyBin.

Wednesday, December 9, 2009

MEMT and METEOR

Kenneth Heafield and Michael Denkowski: Features for System Combination
(This is work done as an MT lab project.)

Michael will give an update on his recent work on the METEOR MT evaluation matrix.

10 Dec 2009, Thursday, 12:00-1:30, in GHC 6501

Monday, November 9, 2009

Lori's talk

Speaker: Lori Levin
Where: GHC 6501
When: Nov 09, 2009 - Tuesday - Noon
Title: A Pendulum Swung Too Far
Abstract:
This paper by Ken Church deals with the never ending battle between Empiricism and Rationalism,
esp. its incarnation in NLP.
Lori will summarize and present the arguments formulated in the
paper. She will then continue with her own views on why linguistics
needs to be brought back into NLP and MT in particular.


Monday, August 10, 2009

Two talks

Talk 1:
Nguyen Bach: Source-side Dependency Tree Reordering Models with Subtree Movements and Constraints

Abstract: We propose a novel source-side dependency tree reordering model for statistical machine translation, in which subtree movements and constraints are represented as reordering events associated with the widely used lexicalized reordering models. This model allows us to not only efficiently capture the statistical distribution of the subtree-to-subtree transitions in training data, but also utilize it directly at the decoding time to guide the search process. Using subtree movements and constraints as features in a log-linear model, we are able to help the reordering models make better selections. It also allows the subtle importance of monolingual syntactic movements to be learned alongside other reordering features. We show improvements in translation quality in English-Spanish and English-Iraqi translation tasks.

This is joint work with Qin Gao and Stephan Vogel.

Talk 2:
Francisco (Paco) Guzman: Reassessment of the Role of Phrase Extraction in SMT

Abstract: In this paper we study in detail the relation between word alignment and phrase extraction. First, we analyze different word alignments according to several characteristics and compare them to hand-aligned data. Then, we analyze the phrase-pairs generated by these alignments. We observed that the number of unaligned words has a large impact on the characteristics of the phrase table. A manual evaluation of phrase pair quality showed that the increase in the number of unaligned words results in a lower quality. Finally, we present translation results from using the number of unaligned words as features from which we obtain up to 2BP of improvement.

This is joint work with Qin Gao and Stephan Vogel.

Monday, June 15, 2009

Making Disfluent Output Slightly Less So: MT System Combination Search Spaces and Optimization

Speaker: Kenneth Heafield

Title: Making Disfluent Output Slightly Less So:
MT System Combination Search Spaces and Optimization

Abstract: System combination merges several machine translation outputs
into a single improved sentence. This talk starts by summarizing the
approach including, a search space derived from the alignments, and
hypothesis scoring. The current search space focuses on picking words
in a roughly word synchronous way. Another search space under development
builds a directed graph in which aligned words correspond to a vertex and
each bigram corresponds to a directed edge. Search is conducted much like
a left-to-right MT decoder. Speed optimizations, which allow decoding at
5.5 sentences per second, apply to other MT systems in the areas of
duplicate handling, language model state, and multithreading. This speed
allows me to find hyperparameters by searching hundreds of parameter
combinations, each with a full round of tuning. In preparation for
last Friday's NIST submission, system combination improved 2.4 BLEU
points over the best component system for Urdu to English translation.

Tuesday, April 28, 2009

EBMT with external word alignment and chunk alignment

Title: EBMT with external word alignment and chunk alignment.

Who: Jae Dong Kim
When: Tuesday May 12, 12:00pm
Where: NSH 3305

Abstract: Since both EBMT and SMT are data driven methods, more accurate word alignment improves system performance in EBMT as in SMT. However, EBMT has focused on finding analogous examples while SMT has achieved plausibly accurate word alignment. For this reason, it is natural that one thinks that EBMT can benefit from using SMT word alignment. In this talk, I am going to talk about our approach to make use of more accurate external word alignment from SMT in our EBMT system. I am also going to talk about my preliminary results with chunk alignment for translation in EBMT.

Monday, April 13, 2009

Language Model Adaptation for Difficult to Translate Phrases

Presenter: Behrang Mohit
Title: Language Model Adaptation for Difficult to Translate Phrases
Date: Tuesday 12:30pm, 14 April 2009

Abstract:
We investigate the idea of adapting language models for phrases that
have poor translation quality. We apply a selective adaptation
criterion which uses a classifier to locate the most difficult phrase
of each source language sentence. A special adapted language model is
constructed for the highlighted phrase. Our adaptation heuristic uses
lexical features of the phrase to locate the relevant parts of the
parallel corpus for language model training. As we vary the
experimental setup by changing the size of the SMT training data, our
adaptation method consistently shows strong improvements over the
baseline systems.
This is a joint work with Frank Liberato and Rebecca Hwa.