Filters








4,639 Hits in 6.0 sec

Results of the Translation Inference Across Dictionaries 2019 Shared Task

Jorge Gracia, Besim Kabashi, Ilan Kernerman, Marta Lanau-Coronas, Dorielle Lonke
2019 Zenodo  
The objective of the Translation Inference Across Dictionaries (TIAD) shared task is to explore and compare methods and techniques that infer translations indirectly between language pairs, based on other  ...  The evaluation of the results was carried out by the organisers against manually compiled language pairs of K Dictionaries.  ...  Acknowledgements We would like to thank Michael Ruppert (University of Erlangen-Nuremberg) for his assistance with the Word2Vec baseline.  ... 
doi:10.5281/zenodo.3555154 fatcat:2yhcyak7lbh3vpp43cmw2n2pii

TIAD 2019 Shared Task: Leveraging Knowledge Graphs with Neural Machine Translation for Automatic Multilingual Dictionary Generation

Daniel Torregrosa, Mihael Arcan, Sina Ahmadi, John P. McCrae
2019 Zenodo  
This paper describes the different proposed approaches to the TIAD 2019 Shared Task, which consisted in the automatic discovery and generation of dictionaries leveraging multilingual knowledge bases.  ...  We present three methods based on graph analysis and neural machine translation and show that we can generate translations without parallel data.  ...  Development Fund, and the European Union's Horizon 2020 research and innovation programme under grant agreement No 731015, ELEXIS -European Lexical Infrastructure.  ... 
doi:10.5281/zenodo.3266899 fatcat:2wcdo7v67vbf5kdbzwtvqvgabu

Translation Inference through Multi-lingual Word Embedding Similarity

Kathrin Donandt, Christian Chiarcos
2019 Zenodo  
This paper describes our contribution to the Shared Task on Translation Inference across Dictionaries (TIAD-2019).  ...  We use the similarity of the word embeddings to predict candidate translations.  ...  The conversion of FreeDict dictionaries into TIAD-TSV data that we used for estimating the prediction threshold was performed in the context of the Research and Innovation Action "Pret-a-LLOD.  ... 
doi:10.5281/zenodo.3555183 fatcat:ky5avcys7vd4lglkyuqw3fpvfi

Graph Exploration and Cross-lingual Word Embeddings for Translation Inference Across Dictionaries

Marta Lanau-Coronas, Jorge Gracia
2020 Zenodo  
This paper describes the participation of two different approaches in the 3rd Translation Inference Across Dictionaries (TIAD 2020) shared task.  ...  The aim of the task is to automatically generate new bilingual dictionaries from existing ones.  ...  In this context, the objective of the Translation Inference Across Dictionaries (TIAD) shared task 3 is to automatically generate new bilingual dictionaries based on known translations contained in this  ... 
doi:10.5281/zenodo.3898278 fatcat:4asperaxhzdflnlksnu6ybu3zy

TIAD Shared Task 2019: Orthonormal Explicit Topic Analysis for Translation Inference across Dictionaries

John P. McCrae
2019 Zenodo  
The task of inferring translations can be achieved by the means of comparable corpora and in this paper we apply explicit topic modelling over comparable corpora to the task of inferring translation candidates  ...  The method proves highly effective at selecting translations with high precision.  ...  Development Fund, and the European Unions Horizon 2020 research and innovation programme under grant agreement No 731015, ELEXIS -European Lexical Infrastructure.  ... 
doi:10.5281/zenodo.3266904 fatcat:hby7ijgy3fdf7jo4madpkhsrsq

Multilingual word translation using auxiliary languages

Hagai Taitelbaum, Gal Chechik, Jacob Goldberger
2019 Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)  
Experiments on a standard multilingual word translation benchmark demonstrate that our model outperforms state of the art results.  ...  The actual translation, however, is still performed as an isolated bilingual task.  ...  The same translation principle can be used in the dictionary refinement step of the mapping training process.  ... 
doi:10.18653/v1/d19-1134 dblp:conf/emnlp/TaitelbaumCG19 fatcat:5yopaknevrbjdb2bvlkqie4r4e

Scalable Cross-Lingual Transfer of Neural Sentence Embeddings

Hanan Aldarmaki, Mona Diab
2019 Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*  
We develop and investigate several crosslingual alignment approaches for neural sentence embedding models, such as the supervised inference classifier, InferSent, and sequential encoder-decoder models.  ...  Our results support representation transfer as a scalable approach for modular cross-lingual alignment of neural sentence embeddings, where we observe better performance compared to joint models in intrinsic  ...  Results The results of the various SDAE models compared with the baselines are shown in Figure 5 .  ... 
doi:10.18653/v1/s19-1006 dblp:conf/starsem/AldarmakiD19 fatcat:oxgda73bpfcsphfh5vdpuhkeiy

Scalable Cross-Lingual Transfer of Neural Sentence Embeddings [article]

Hanan Aldarmaki, Mona Diab
2019 arXiv   pre-print
We develop and investigate several cross-lingual alignment approaches for neural sentence embedding models, such as the supervised inference classifier, InferSent, and sequential encoder-decoder models  ...  Our results support representation transfer as a scalable approach for modular cross-lingual alignment of neural sentence embeddings, where we observe better performance compared to joint models in intrinsic  ...  Results The results of the various SDAE models compared with the baselines are shown in Figure 5 .  ... 
arXiv:1904.05542v1 fatcat:cvmxdq7kvzbhphsanx5rlq6rbi

Emerging Cross-lingual Structure in Pretrained Language Models [article]

Shijie Wu and Alexis Conneau and Haoran Li and Luke Zettlemoyer and Veselin Stoyanov
2020 arXiv   pre-print
The only requirement is that there are some shared parameters in the top layers of the multi-lingual encoder.  ...  We show, contrary to what was previously hypothesized, that transfer is possible even when there is no shared vocabulary across the monolingual corpora and also when the text comes from very different  ...  For each training word w 1 in the bilingual dictionary, we either leave it as is (70% of the time) or randomly replace it with one of the possible translations from the dictionary (30% of the time).  ... 
arXiv:1911.01464v3 fatcat:zcutmu7pq5hyxmpirkf243v6jy

Cross-lingual Alignment Methods for Multilingual BERT: A Comparative Study [article]

Saurabh Kulshreshtha, José Luis Redondo-García, Ching-Yun Chang
2020 arXiv   pre-print
We evaluate the performance of different alignment methodologies across eight languages on two tasks: Name Entity Recognition and Semantic Slot Filling.  ...  Importantly we identify the biases of the alignment methods to the type of task and proximity to the transfer language.  ...  We compare the effect of dictionary and parallel corpora supervision across all tasks.  ... 
arXiv:2009.14304v1 fatcat:ultegtidpbaipj3r6sb4c7rbku

CoSDA-ML: Multi-Lingual Code-Switching Data Augmentation for Zero-Shot Cross-Lingual NLP [article]

Libo Qin, Minheng Ni, Yue Zhang, Wanxiang Che
2020 arXiv   pre-print
Experimental results on five tasks with 19 languages show that our method leads to significantly improved performances for all the tasks compared with mBERT.  ...  However, these models are limited by having inconsistent contextualized representations of subwords across different languages.  ...  Acknowledgements This work was supported by the National Natural Science Foundation of China (NSFC) via grant 61976072, 61632011 and 61772153.  ... 
arXiv:2006.06402v2 fatcat:klduee3cgzdftnnjtjke4hfwti

Improving Precision of Grammatical Error Correction with a Cheat Sheet

Mengyang Qiu, Xuejiao Chen, Maggie Liu, Krishna Parvathala, Apurva Patil, Jungyeul Park
2019 Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications  
BEA 2019 Shared Task on GEC.  ...  Our results show that phrases directly extracted from GEC corpora outperform phrases from a statistical machine translation phrase table by a large margin.  ...  During inference, we used a beam size of 10. Table 2 shows the baseline result and the results we submitted to the BEA 2019 Shared Task.  ... 
doi:10.18653/v1/w19-4425 dblp:conf/bea/QiuCLPPP19 fatcat:qxvvmf4oqfclfn42cnwetv2niq

Cross-lingual Alignment vs Joint Training: A Comparative Study and A Simple Unified Framework [article]

Zirui Wang, Jiateng Xie, Ruochen Xu, Yiming Yang, Graham Neubig, Jaime Carbonell
2020 arXiv   pre-print
Our empirical results reveal a set of pros and cons for both methods, and show that the relative performance of alignment versus joint training is task-dependent.  ...  In this paper, we first conduct direct comparisons of representations learned using both of these methods across diverse cross-lingual tasks.  ...  The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the U.S. government.  ... 
arXiv:1910.04708v4 fatcat:7fld5meanfftnikjsqfdfbhu5i

Context-Aware Cross-Lingual Mapping [article]

Hanan Aldarmaki, Mona Diab
2019 arXiv   pre-print
Cross-lingual word vectors are typically obtained by fitting an orthogonal matrix that maps the entries of a bilingual dictionary from a source to a target vector space.  ...  In our experiments, both approaches resulted in cross-lingual sentence embeddings that outperformed context-independent word mapping in sentence translation retrieval.  ...  Since this can result in a very large dictionary, we capped the number of dictionary words at 1M for efficiency.  ... 
arXiv:1903.03243v2 fatcat:bgy3m2lhurdrjkhcojmgiu5pjm

Context-Aware Cross-Lingual Mapping

Hanan Aldarmaki, Mona Diab
2019 Proceedings of the 2019 Conference of the North  
Cross-lingual word vectors are typically obtained by fitting an orthogonal matrix that maps the entries of a bilingual dictionary from a source to a target vector space.  ...  In our experiments, both approaches resulted in cross-lingual sentence embeddings that outperformed context-independent word mapping in sentence translation retrieval.  ...  Since this can result in a very large dictionary, we capped the number of dictionary words at 1M for efficiency.  ... 
doi:10.18653/v1/n19-1391 dblp:conf/naacl/AldarmakiD19 fatcat:wzw7ssx3over7h3uk6heh6siwu
« Previous Showing results 1 — 15 out of 4,639 results