@inproceedings{abdelali-durrani-guzman:2016:N16-3, address = {San Diego, California}, author = {Abdelali, Ahmed and Durrani, Nadir and Guzm\'{a}n, Francisco}, booktitle = {Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Demonstrations}, link = {http://www.aclweb.org/anthology/N16-3004}, month = {June}, pages = {17--21}, publisher = {Association for Computational Linguistics}, title = {iAppraise: A Manual Machine Translation Evaluation Environment Supporting Eye-tracking}, year = {2016} }

@inproceedings{guzman-nakov-marquez:2016:SemEval, address = {San Diego, California}, author = {Guzm\'{a}n, Francisco and Nakov, Preslav and M\`{a}rquez, Llu\'{i}s}, booktitle = {Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016)}, link = {http://www.aclweb.org/anthology/S16-1137}, month = {June}, pages = {887--895}, publisher = {Association for Computational Linguistics}, title = {MTE-NN at SemEval-2016 Task 3: Can Machine Translation Evaluation Help Community Question Answering?}, year = {2016} }

@inproceedings{zaghouani-EtAl:2016:SeMaT, address = {Austin, Texas}, author = {Zaghouani, Wajdi and Abdelali, Ahmed and Guzm\'{a}n, Francisco and Sajjad, Hassan}, booktitle = {Proceedings of the Workshop on Semitic Machine Translation}, link = {http://www.aclweb.org/anthology/W/W05/W05-0204}, month = {November}, pages = {20--27}, publisher = {Association for Computational Linguistics}, title = {Normalizing Mathematical Expressions to Improve the Translation of Educational Content}, year = {2016} }

Evaluation of machine translation (MT) into morphologically rich languages (MRL) has not been well studied despite posing many challenges. In this paper, we explore the use of embeddings obtained from different levels of lexical and morpho-syntactic linguistic analysis and show that they improve MT evaluation into an MRL. Specifically we report on Arabic, a language with complex and rich morphology. Our results show that using a neural-network model with different input representations produces results that clearly outperform the state-of-the-art for MT evaluation into Arabic, by almost over 75% increase in correlation with human judgments on pairwise MT evaluation quality task. More importantly, we demonstrate the usefulness of morpho-syntactic representations to model sentence similarity for MT evaluation and address complex linguistic phenomena of Arabic.

@inproceedings{guzman-EtAl:2016:COLING, abstract = {Evaluation of machine translation (MT) into morphologically rich languages (MRL) has not been well studied despite posing many challenges. In this paper, we explore the use of embeddings obtained from different levels of lexical and morpho-syntactic linguistic analysis and show that they improve MT evaluation into an MRL. Specifically we report on Arabic, a language with complex and rich morphology. Our results show that using a neural-network model with different input representations produces results that clearly outperform the state-of-the-art for MT evaluation into Arabic, by almost over 75% increase in correlation with human judgments on pairwise MT evaluation quality task. More importantly, we demonstrate the usefulness of morpho-syntactic representations to model sentence similarity for MT evaluation and address complex linguistic phenomena of Arabic.}, address = {Osaka, Japan}, author = {Guzm\'{a}n, Francisco and Bouamor, Houda and Baly, Ramy and Habash, Nizar}, booktitle = {Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers}, link = {http://aclweb.org/anthology/C16-1132}, month = {December}, pages = {1398--1408}, publisher = {The COLING 2016 Organizing Committee}, title = {Machine Translation Evaluation for Arabic using Morphologically-enriched Embeddings}, year = {2016} }

The paper describes the Egyptian Arabic-to-English statistical machine translation (SMT) system that the QCRI-Columbia-NYUAD (QCN) group submitted to the NIST OpenMT'2015 competition. The competition focused on informal dialectal Arabic, as used in SMS, chat, and speech. Thus, our efforts focused on processing and standardizing Arabic, e.g., using tools such as 3arrib and MADAMIRA. We further trained a phrase-based SMT system using state-of-the-art features and components such as operation sequence model, class-based language model, sparse features, neural network joint model, genre-based hierarchically-interpolated language model, unsupervised transliteration mining, phrase-table merging, and hypothesis combination. Our system ranked second on all three genres.

@Proceedings{sajjad2015-NIST, title={ The QCN Egyptian Arabic to English Statistical Machine Translation System for NIST OpenMT’2015}, author={Sajjad, Hassan and Durrani, Nadir and Guzmán, Francisco and Nakov, Preslav and Abdelali, Ahmed and Vogel, Stephan and Salloum, Wael and El Kholy, Ahmed and Habash, Nizar}, year={2015}, abstract={The paper describes the Egyptian Arabic-to-English statistical machine translation (SMT) system that the QCRI-Columbia-NYUAD (QCN) group submitted to the NIST OpenMT'2015 competition. The competition focused on informal dialectal Arabic, as used in SMS, chat, and speech. Thus, our efforts focused on processing and standardizing Arabic, e.g., using tools such as 3arrib and MADAMIRA. We further trained a phrase-based SMT system using state-of-the-art features and components such as operation sequence model, class-based language model, sparse features, neural network joint model, genre-based hierarchically-interpolated language model, unsupervised transliteration mining, phrase-table merging, and hypothesis combination. Our system ranked second on all three genres.}, booktitle={Proceedings of the NIST Open Machine Translation Evaluation Workshop}, publisher={NIST} }

QAT2 is a multimedia content translation web service developed by QCRI to help content provider to reach audiences and viewers speaking different languages. It is built with open source technologies such as KALDI, Moses and MaryTTS, to provide a complete translation experience for web users. It translates text content in its original format, and produce translated videos with speech-to-speech translation. The result is a complete native language experience for end users on foreign language websites. The system currently supports translation from Arabic to English.

@inproceedings{abdelali2015-INTERSPEECH, title={{QAT}$^2$--The {QCRI} Advanced Transcription and Translation System}, author={Abdelali, Ahmed and Ali, Ahmed and Guzm{\'a}n, Francisco and Stahlberg, Felix and Vogel, Stephan and Zhang, Yifan}, booktitle={ Proceedings of the 16th Annual Conference of the International Speech Communication Association}, year={2015}, abstract={QAT2 is a multimedia content translation web service developed by QCRI to help content provider to reach audiences and viewers speaking different languages. It is built with open source technologies such as KALDI, Moses and MaryTTS, to provide a complete translation experience for web users. It translates text content in its original format, and pro- duce translated videos with speech-to-speech translation. The result is a complete native language experience for end users on foreign language websites. The system currently supports translation from Arabic to English.} }

We study the impact of source length and verbosity of the tuning dataset on the performance of parameter optimizers such as MERT and PRO for statistical machine translation. In particular, we test whether the verbosity of the resulting translations can be modified by varying the length or the verbosity of the tuning sentences. We find that MERT learns the tuning set verbosity very well, while PRO is sensitive to both the verbosity and the length of the source sentences in the tuning set; yet, overall PRO learns best from high-verbosity tuning datasets. Given these dependencies, and potentially some other such as amount of reordering, number of unknown words, syntactic complexity, and evaluation measure, to mention just a few, we argue for the need of controlled evaluation scenarios, so that the selection of tuning set and optimization strategy does not overshadow scientific advances in modeling or decoding. In the mean time, until we develop such controlled scenarios, we recommend using PRO with a large verbosity tuning set, which, in our experiments, yields highest BLEU across datasets and language pairs.

@InProceedings{guzman-nakov-vogel:2015:CoNLL, author = {Guzm\'{a}n, Francisco and Nakov, Preslav and Vogel, Stephan}, title = {Analyzing Optimization for Statistical Machine Translation: MERT Learns Verbosity, PRO Learns Length}, booktitle = {Proceedings of the Nineteenth Conference on Computational Natural Language Learning}, month = {July}, year = {2015}, address = {Beijing, China}, publisher = {Association for Computational Linguistics}, pages = {62--72}, url = {http://www.aclweb.org/anthology/K15-1007}, abstract = {We study the impact of source length and verbosity of the tuning dataset on the performance of parameter optimizers such as MERT and PRO for statistical machine translation. In particular, we test whether the verbosity of the resulting translations can be modified by varying the length or the verbosity of the tuning sentences. We find that MERT learns the tuning set verbosity very well, while PRO is sensitive to both the verbosity and the length of the source sentences in the tuning set; yet, overall PRO learns best from high-verbosity tuning datasets. Given these dependencies, and potentially some other such as amount of reordering, number of unknown words, syntactic complexity, and evaluation measure, to mention just a few, we argue for the need of controlled evaluation scenarios, so that the selection of tuning set and optimization strategy does not overshadow scientific advances in modeling or decoding. In the mean time, until we develop such controlled scenarios, we recommend using PRO with a large verbosity tuning set, which, in our experiments, yields highest BLEU across datasets and language pairs.} }

This paper presents the AMARA corpus of on-line educational content: a new parallel corpus of educational video subtitles, multilingually aligned for 20 languages, i.e. 20 monolingual corpora and 190 parallel corpora. This corpus includes both resource-rich languages such as English and Arabic, and resource-poor languages such as Hindi and Thai. In this paper, we describe the gathering, validation, and preprocessing of a large collection of parallel, community-generated subtitles. Furthermore, we describe the methodology used to prepare the data for Machine Translation tasks. Additionally, we provide a document-level, jointly aligned development and test sets for 14 language pairs, designed for tuning and testing Machine Translation systems. We provide baseline results for these tasks, and highlight some of the challenges we face when building machine translation systems for educational content.

@inproceedings{abdelali:2014:amara, abstract = {This paper presents the AMARA corpus of on-line educational content: a new parallel corpus of educational video subtitles, multilingually aligned for 20 languages, i.e. 20 monolingual corpora and 190 parallel corpora. This corpus includes both resource-rich languages such as English and Arabic, and resource-poor languages such as Hindi and Thai. In this paper, we describe the gathering, validation, and preprocessing of a large collection of parallel, community-generated subtitles. Furthermore, we describe the methodology used to prepare the data for Machine Translation tasks. Additionally, we provide a document-level, jointly aligned development and test sets for 14 language pairs, designed for tuning and testing Machine Translation systems. We provide baseline results for these tasks, and highlight some of the challenges we face when building machine translation systems for educational content.}, address = {Reykjavik, Iceland}, author = {Abdelali, Ahmed and Guzm{\'a}n, Francisco and Sajjad, Hassan and Vogel, Stephan}, booktitle = {Proceedings of the Ninth International Conference on Language Resources and Evaluation ({LREC}'14)}, month = {april}, title = {The {AMARA} Corpus: Building parallel language resources for the educational domain}, year = {2014} }

In this paper, we present the main features of the Amara project, and its impact on the accessibility landscape with the use of innovative technology. We also show the effectiveness of volunteer communities in addressing large subtitling and translation tasks, that accompany the ever-growing amounts of online video content. Furthermore, we present two different applications for the platform. First, we examine the growing interest of organizations to build their own subtitling communities. Second, we present how the community-generated material can be used to advance the state-of-the-art of research in fields such as Statistical Machine Translation with focus on educational translation. We provide examples on how both tasks can be achieved successfully.

@inproceedings{jansen:2014:amara, abstract = {In this paper, we present the main features of the Amara project, and its impact on the accessibility landscape with the use of innovative technology. We also show the effectiveness of volunteer communities in addressing large subtitling and translation tasks, that accompany the ever-growing amounts of online video content. Furthermore, we present two different applications for the platform. First, we examine the growing interest of organizations to build their own subtitling communities. Second, we present how the community-generated material can be used to advance the state-of-the-art of research in fields such as Statistical Machine Translation with focus on educational translation. We provide examples on how both tasks can be achieved successfully.}, address = {Heraklion, Greece}, author = {Jansen, Dean and Alcala, Aleli and Guzm{\'a}n, Francisco}, booktitle = {Universal Access in Human-Computer Interaction. Design for All and Accessibility Practice}, month = {June}, pages = {401--411}, publisher = {Springer International Publishing}, title = {Amara: A Sustainable, Global Solution for Accessibility, Powered by Communities of Volunteers}, year = {2014} }

We present novel automatic metrics for machine translation evaluation that use discourse structure and convolution kernels to compare the discourse tree of an automatic translation with that of the human reference. We experiment with five transformations and augmentations of a base discourse tree representation based on the rhetorical structure theory, and we combine the kernel scores for each of them into a single score. Finally, we add other metrics from the ASIYA MT evaluation toolkit, and we tune the weights of the combination on actual human judgments. Experiments on the WMT12 and WMT13 metrics shared task datasets show correlation with human judgments that outperforms what the best systems that participated in these years achieved, both at the segment and at the system level.

@inproceedings{joty-EtAl:2014:W14-33, abstract = {We present novel automatic metrics for machine translation evaluation that use discourse structure and convolution kernels to compare the discourse tree of an automatic translation with that of the human reference. We experiment with five transformations and augmentations of a base discourse tree representation based on the rhetorical structure theory, and we combine the kernel scores for each of them into a single score. Finally, we add other metrics from the ASIYA MT evaluation toolkit, and we tune the weights of the combination on actual human judgments. Experiments on the WMT12 and WMT13 metrics shared task datasets show correlation with human judgments that outperforms what the best systems that participated in these years achieved, both at the segment and at the system level.}, address = {Baltimore, Maryland, USA}, author = {Joty, Shafiq and Guzm\'{a}n, Francisco and M\`{a}rquez, Llu\'{i}s and Nakov, Preslav}, booktitle = {Proceedings of the Ninth Workshop on Statistical Machine Translation ({WMT}'14)}, link = {http://www.aclweb.org/anthology/W/W14/W14-3352}, month = {June}, pages = {402--408}, publisher = {Association for Computational Linguistics}, title = {DiscoTK: Using Discourse Structure for Machine Translation Evaluation}, year = {2014} }

Research on statistical machine translation has focused on particular translation directions, typically with English as the target language, e.g., from Arabic to English. When we reverse the translation direction, the multiple reference translations turn into multiple possible inputs, which offers both challenges and opportunities. We propose and evaluate several strategies for making use of these multiple inputs: (a) select one of the datasets, (b) select the best input for each sentence, and (c) synthesize an input for each sentence by fusing the available inputs. Surprisingly, we find out that it is best to tune on the hardest available input, not on the one that yields the highest BLEU score. This finding has implications on how to pick good translators and how to select useful data for parameter optimization in SMT.

@inproceedings{nakov:2013:parameter, abstract = {Research on statistical machine translation has focused on particular translation directions, typically with English as the target language, e.g., from Arabic to English. When we reverse the translation direction, the multiple reference translations turn into multiple possible inputs, which offers both challenges and opportunities. We propose and evaluate several strategies for making use of these multiple inputs: (a) select one of the datasets, (b) select the best input for each sentence, and (c) synthesize an input for each sentence by fusing the available inputs. Surprisingly, we find out that it is best to tune on the hardest available input, not on the one that yields the highest BLEU score. This finding has implications on how to pick good translators and how to select useful data for parameter optimization in SMT.}, author = {Nakov, Preslav and Al Obaidli, Fahad and Guzm{\'a}n, Francisco and Vogel, Stephan}, booktitle = {Proceedings of the International Conference Recent Advances in Natural Language Processing ({RANLP}'13)}, month = {September}, title = {Parameter Optimization for Statistical Machine Translation: It Pays to Learn from Hard Examples}, year = {2013} }

We describe the Arabic-English and English-Arabic statistical machine translation systems developed by the Qatar Computing Research Institute for the IWSLT’2013 evaluation campaign on spoken language translation. We used one phrase-based and two hierarchical decoders, exploring various settings thereof. We further experimented with three domain adaptation methods, and with various Arabic word segmentation schemes. Combining the output of several systems yielded a gain of up to 3.4 BLEU points over the baseline. Here we also describe a specialized normalization scheme for evaluating Arabic output, which was adopted for the IWSLT’2013 evaluation campaign.

@inproceedings{sajjad:2013:qcri, abstract = {We describe the Arabic-English and English-Arabic statistical machine translation systems developed by the Qatar Computing Research Institute for the IWSLT’2013 evaluation campaign on spoken language translation. We used one phrase-based and two hierarchical decoders, exploring various settings thereof. We further experimented with three domain adaptation methods, and with various Arabic word segmentation schemes. Combining the output of several systems yielded a gain of up to 3.4 BLEU points over the baseline. Here we also describe a specialized normalization scheme for evaluating Arabic output, which was adopted for the IWSLT’2013 evaluation campaign.}, address = {Heidelberg, Germany}, author = {Sajjad, Hassan and Guzm{\'a}n, Francisco and Nakov, Preslav and Abdelali, Ahmed and Murray, Kenton and Al Obaidli, Fahad and Vogel, Stephan}, booktitle = {Proceedings of the 10th International Workshop on Spoken Language Translation {(IWSLT'13)}}, month = {December}, title = {{QCRI} at {IWSLT} 2013: Experiments in Arabic-English and English-Arabic Spoken Language Translation}, volume = {13}, year = {2013} }

We describe the systems developed by the team of the Qatar Computing Research Institute for the WMT12 Shared Translation Task. We used a phrase-based statistical machine translation model with several non-standard settings, most notably tuning data selection and phrase table combination. The evaluation results show that we rank second in BLEU and TER for Spanish-English, and in the top tier for German-English.

@InProceedings{guzman-EtAl:2012:WMT, author = {Guzm{\'a}n, Francisco and Nakov, Preslav and Thabet, Ahmed and Vogel, Stephan}, title = {{QCRI} at {WMT}12: Experiments in Spanish-English and German-English Machine Translation of News Text}, booktitle = {Proceedings of the Seventh Workshop on Statistical Machine Translation ({WMT'12})}, month = {June}, year = {2012}, address = {Montr{\'e}al, Canada}, publisher = {Association for Computational Linguistics}, pages = {298--303}, url = {http://www.aclweb.org/anthology/W12-3136}, abstract = {We describe the systems developed by the team of the Qatar Computing Research Institute for the WMT12 Shared Translation Task. We used a phrase-based statistical machine translation model with several non-standard settings, most notably tuning data selection and phrase table combination. The evaluation results show that we rank second in BLEU and TER for Spanish-English, and in the top tier for German-English.} }

abstract = {We present a framework for the analysis of Machine Translation performance. We use multivariate linear models to determine the impact of a wide range of features on translation performance. Our assumption is that variables that most contribute to predict translation performance are the key to understand the differences between good and bad translations. During training, we learn the regression parameters that better predict translation quality using a wide range of input features based on the translation model and the first-best translation hypotheses. We use a linear regression with regularization. Our results indicate that with regularized linear regression, we can achieve higher levels of correlation between our predicted values and the actual values of the quality metrics. Our analysis shows that the performance for in-domain data is largely dependent on the characteristics of the translation model. On the other hand, out-of domain data can benefit from better reordering strategies.

@InProceedings{guzman-vogel:2012:PAPERS, author = {Guzm{\'a}n, Francisco and Vogel, Stephan}, title = {Understanding the Performance of Statistical {MT} Systems: A Linear Regression Framework}, booktitle = {Proceedings of the 24rd International Conference on Computational Linguistics {(COLING 2012)}, month = {December}, year = {2012}, address = {Mumbai, India}, pages = {1029--1044}, url = {http://www.aclweb.org/anthology/C12-1063}, abstract = {We present a framework for the analysis of Machine Translation performance. We use multivariate linear models to determine the impact of a wide range of features on translation performance. Our assumption is that variables that most contribute to predict translation performance are the key to understand the differences between good and bad translations. During training, we learn the regression parameters that better predict translation quality using a wide range of input features based on the translation model and the first-best translation hypotheses. We use a linear regression with regularization. Our results indicate that with regularized linear regression, we can achieve higher levels of correlation between our predicted values and the actual values of the quality metrics. Our analysis shows that the performance for in-domain data is largely dependent on the characteristics of the translation model. On the other hand, out-of domain data can benefit from better reordering strategies. }

Statistical Word Alignments represent lexical word-to-word translations between source and target language sentences. They are considered the starting point for many state of the art Statistical Machine Translation (SMT) systems. In phrase-based systems, word alignments are loosely linked to the translation model. Despite the improvements reached in word alignment quality, there has been a modest improvement in the end-to-end translation. Until recently, little or no attention was paid to the structural characteristics of word-alignments (e.g. unaligned words) and their impact in further stages of the phrase-based SMT pipeline. A better understanding of the relationship between word alignment and the entailing processes will help to identify the variables across the pipeline that most influence translation performance and can be controlled by modifying word alignment’s characteristics. In this dissertation, we perform an in-depth study of the impact of word alignments at different stages of the phrase-based statistical machine translation pipeline, namely word alignment, phrase extraction, phrase scoring and decoding. Moreover, we establish a multivariate prediction model for different variables of word alignments, phrase tables and translation hypotheses. Based on those models, we identify the most important alignment variables and propose two alternatives to provide more control over alignment structure and thus improve SMT. Our results show that using alignment structure into decoding, via alignment gap features yields significant improvements, specially in situations where translation data is limited. During the development of this dissertation we discovered how different characteristics of the alignment impact Machine Translation. We observed that while good quality alignments yield good phrase-pairs, the consolidation of a translation model is dependent on the alignment structure, not quality. Human-alignments are more dense than the computer generated counterparts, which trend to be more sparse and precision-oriented. Trying to emulate human-like alignment structure resulted in poorer systems, because the resulting translation models trend to be more compact and lack translation options. On the other hand, more translation options, even if they are noisier, help to improve the quality of the translation. This is due to the fact that translation does not rely only on the translation model, but also other factors that help to discriminate the noise from bad translations (e.g. the language model). Lastly, when we provide the decoder with features that help it to make more `informed decisions'' we observe a clear improvement in translation quality. This was specially true for the discriminative alignments which inherently leave more unaligned words. The result is more evident in low-resource settings where having larger translation lexicons represent more translation options. Using simple features to help the decoder discriminate translation hypotheses, clearly showed consistent improvements.

@phdthesis{guzman-thesis:2011, abstract = {Statistical Word Alignments represent lexical word-to-word translations between source and target language sentences. They are considered the starting point for many state of the art Statistical Machine Translation (SMT) systems. In phrase-based systems, word alignments are loosely linked to the translation model. Despite the improvements reached in word alignment quality, there has been a modest improvement in the end-to-end translation. Until recently, little or no attention was paid to the structural characteristics of word-alignments (e.g. unaligned words) and their impact in further stages of the phrase-based SMT pipeline. A better understanding of the relationship between word alignment and the entailing processes will help to identify the variables across the pipeline that most influence translation performance and can be controlled by modifying word alignment’s characteristics. In this dissertation, we perform an in-depth study of the impact of word alignments at different stages of the phrase-based statistical machine translation pipeline, namely word alignment, phrase extraction, phrase scoring and decoding. Moreover, we establish a multivariate prediction model for different variables of word alignments, phrase tables and translation hypotheses. Based on those models, we identify the most important alignment variables and propose two alternatives to provide more control over alignment structure and thus improve SMT. Our results show that using alignment structure into decoding, via alignment gap features yields significant improvements, specially in situations where translation data is limited. During the development of this dissertation we discovered how different characteristics of the alignment impact Machine Translation. We observed that while good quality alignments yield good phrase-pairs, the consolidation of a translation model is dependent on the alignment structure, not quality. Human-alignments are more dense than the computer generated counterparts, which trend to be more sparse and precision-oriented. Trying to emulate human-like alignment structure resulted in poorer systems, because the resulting translation models trend to be more compact and lack translation options. On the other hand, more translation options, even if they are noisier, help to improve the quality of the translation. This is due to the fact that translation does not rely only on the translation model, but also other factors that help to discriminate the noise from bad translations (e.g. the language model). Lastly, when we provide the decoder with features that help it to make more `informed decisions'' we observe a clear improvement in translation quality. This was specially true for the discriminative alignments which inherently leave more unaligned words. The result is more evident in low-resource settings where having larger translation lexicons represent more translation options. Using simple features to help the decoder discriminate translation hypotheses, clearly showed consistent improvements.}, author = {Guzm{\'a}n, Francisco}, month = {December}, school = {Instituto Tecnológico y de Estudios Superiores de Monterrey, Campus Monterrey}, title = {The Impact of Statistical Word Alignment Quality and Structure in Phrase Based Statistical Machine Translation}, year = {2011} }

This paper proposes a novel semi-supervised word alignment technique called EMDC that integrates discriminative and generative methods. A discriminative aligner is used to find high precision partial alignments that serve as constraints for a generative aligner which implements a constrained version of the EM algorithm. Experiments on small-size Chinese and Arabic tasks show consistent improvements on AER. We also experimented with moderate-size Chinese machine translation tasks and got an average of 0.5 point improvement on BLEU scores across five standard NIST test sets and four other test sets.

@inproceedings{gao:2010:emdc, Author = {Gao, Qin and Guzm{\'a}n, Francisco and Vogel, Stephan}, Booktitle = {Proceedings of the 23rd International Conference on Computational Linguistics {(COLING 2010)}}, Organization = {Association for Computational Linguistics}, Pages = {349--357}, Title = {EMDC: a semi-supervised approach for word alignment}, address = {Beijing, China}, Year = {2010}, Abstract= This paper proposes a novel semi-supervised word alignment technique called EMDC that integrates discriminative and generative methods. A discriminative aligner is used to find high precision partial alignments that serve as constraints for a generative aligner which implements a constrained version of the EM algorithm. Experiments on small-size Chinese and Arabic tasks show consistent improvements on AER. We also experimented with moderate-size Chinese machine translation tasks and got an average of 0.5 point improvement on BLEU scores across five standard NIST test sets and four other test sets.}

In this paper we present an analysis of a phrase-based machine translation methodology that integrates paraphrases obtained from an intermediary language (French) for translations between Spanish and English. The purpose of the research presented in this document is to find out how much extra information (i.e. improvements in translation quality) can be found when using Translation Paraphrases (TPs). In this document we present an extensive statistical analysis to support conclusions.

@inproceedings{guzman:2008:translation, Author = {Guzm{\'a}n, Francisco and Garrido, Leonardo}, Booktitle = {Computational Linguistics and Intelligent Text Processing {(CICLing'08)}}, Pages = {388--398}, address = {Haifa, Israel}, Title = {Translation paraphrases in phrase-based machine translation}, Publisher = {Springer Berlin Heidelberg}, Month = { February}, Year = {2008}, Abstract = { In this paper we present an analysis of a phrase-based machine translation methodology that integrates paraphrases obtained from an intermediary language (French) for translations between Spanish and English. The purpose of the research presented in this document is to find out how much extra information (i.e. improvements in translation quality) can be found when using Translation Paraphrases (TPs). In this document we present an extensive statistical analysis to support conclusions.}}

Statistical methods have proven to be very effective when addressing linguistic problems, specially when dealing with Machine Translation. Nevertheless, Statistical Machine Translation effectiveness is limited to situations where large amounts of training data are available. Therefore, the broader the coverage of a SMT system is, the better the chances to get a reasonable output are. In this paper we propose a method to improve quality of translations of a phrase-based Machine Translation system by extending phrase-tables with the use of translation paraphrases learned from a third language. Our experiments were done translating from Spanish to English pivoting through French.

@inproceedings{guzman:2007:using, Author = {Guzm{\'a}n, Francisco and Garrido, Leonardo}, Booktitle = {Sixth Mexican International Conference on Artificial Intelligence {(MICAI'07)}. }, Organization = {IEEE}, Pages = {163--172}, Title = {Using Translation Paraphrases from Trilingual Corpora to Improve Phrase-Based Statistical Machine Translation: A Preliminary Report}, Month = {November}, address = { Aguascalientes, Mexico}, Year = {2007}, Abstract = {Statistical methods have proven to be very effective when addressing linguistic problems, specially when dealing with Machine Translation. Nevertheless, Statistical Machine Translation effectiveness is limited to situations where large amounts of training data are available. Therefore, the broader the coverage of a SMT system is, the better the chances to get a reasonable output are. In this paper we propose a method to improve quality of translations of a phrase-based Machine Translation system by extending phrase-tables with the use of translation paraphrases learned from a third language. Our experiments were done translating from Spanish to English pivoting through French.}}

Present day traffic networks are unable to efficiently handle the daily car traffic through urban areas. We think that multiagent systems are an excellent way of doing microscopic simulation and thus provide possible solutions to the traffic control problem. In this paper, we present our simulation-based study to simulate traffic networks and optimize them via a multiagent cooperative system for traffic light control. This system simulates the traffic on an intersection, minimizing the time that each car has to wait in order to be served. Light agents can communicate each other in order to negotiate and share their light times. Our experimental results have shown how our approach can decrease the average car delay while the spawn probability is increased varying the service time and the number of traffic lights sets at a specific intersection. These results show important improvements using our multiagent light control system.

@inproceedings{guzman:2005:towards, Author = {Guzm{\'a}n, Francisco and Garrido, Leonardo}, Booktitle = {Proceedings of the 2005 Agent--Directed Simulation Symposium ({ADS05}) at the 2005 Spring Simulation Multiconference ({SpringSim'05})}, Pages = {29--35}, Title = {Towards Traffic Light Control Through a Cooperative Multiagent System: A Simulation--Based Study}, address = { San Diego, California, USA}, month = { April}, Year = {2005}, Abstract = {Present day traffic networks are unable to efficiently handle the daily car traffic through urban areas. We think that multiagent systems are an excellent way of doing microscopic simulation and thus provide possible solutions to the traffic control problem. In this paper, we present our simulation-based study to simulate traffic networks and optimize them via a multiagent cooperative system for traffic light control. This system simulates the traffic on an intersection, minimizing the time that each car has to wait in order to be served. Light agents can communicate each other in order to negotiate and share their light times. Our experimental results have shown how our approach can decrease the average car delay while the spawn probability is increased varying the service time and the number of traffic lights sets at a specific intersection. These results show important improvements using our multiagent light control system.}}

@techreport{guzman:2005:multiagent, Author = {Guzm{\'a}n, Francisco and Garrido, Leonardo}, Institution = {Tecnol{\'o}gico de Monterrey ({ITESM}), Monterrey, NL, M{\'e}xico}, Title = {Multiagent--Based Traffic Simulation}, Journal = {Technical Report CSI-RI-005}, address = { Monterrey, Mexico}, Year = {2005}}