Filters








69,785 Hits in 5.8 sec

A translation model for sentence retrieval

Vanessa Murdock, W. Bruce Croft
2005 Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing - HLT '05   unpublished
In this work we propose a translation model for monolingual sentence retrieval. We propose four methods for constructing a parallel corpus.  ...  Further, we demonstrate that smoothing from the local context of the sentence improves retrieval over the query likelihood baseline.  ...  This work was supported in part by the Center for Intelligent Information Retrieval, in part by Advanced Research and Development Activity and NSF grant #CCF-0205575 , and in part by SPAWARSYSCEN-SD grant  ... 
doi:10.3115/1220575.1220661 fatcat:r4iuilkz2rgwvh3nlnophm32rm

Guiding Neural Machine Translation with Retrieved Translation Pieces

Jingyi Zhang, Masao Utiyama, Eiichro Sumita, Graham Neubig, Satoshi Nakamura
2018 Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers)  
Specifically, for an input sentence, we use a search engine to retrieve sentence pairs whose source sides are similar with the input sentence, and then collect n-grams that are both in the retrieved target  ...  We compute pseudoprobabilities for each retrieved sentence based on similarities between the input sentence and the retrieved source sentences, and use these to weight the retrieved translation pieces.  ...  Acknowledgments We thank Jiatao Gu for providing their preprocessed datasets in Section 4.5.  ... 
doi:10.18653/v1/n18-1120 dblp:conf/naacl/ZhangUSNN18 fatcat:4wrpgdku45hr7nxtqqzf57kjcu

Guiding Neural Machine Translation with Retrieved Translation Pieces [article]

Jingyi Zhang, Masao Utiyama, Eiichro Sumita, Graham Neubig, Satoshi Nakamura
2018 arXiv   pre-print
Specifically, for an input sentence, we use a search engine to retrieve sentence pairs whose source sides are similar with the input sentence, and then collect n-grams that are both in the retrieved target  ...  We compute pseudo-probabilities for each retrieved sentence based on similarities between the input sentence and the retrieved source sentences, and use these to weight the retrieved translation pieces  ...  Acknowledgments We thank Jiatao Gu for providing their preprocessed datasets in Section 4.5.  ... 
arXiv:1804.02559v1 fatcat:qwouupto6rdcxamqkvokee6hou

A comparison of sentence retrieval techniques

Niranjan Balasubramanian, James Allan, W. Bruce Croft
2007 Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval - SIGIR '07  
Acknowledgments This work was supported in part by the Center for Intelligent Information Retrieval, in part by the Defense Advanced Research Projects Agency (DARPA) under contract number HR0011-06-C-0023  ...  Translation based Models Murdock [8] showed that Model-S, a translation based model, is effective for sentence retrieval, especially for question answering and novelty detection tasks.  ...  Jeon et al [4] describe a mixture model of query likelihood and a translation based model for successfully identifying similar questions.  ... 
doi:10.1145/1277741.1277922 dblp:conf/sigir/BalasubramanianAC07 fatcat:xmva7dwetnh77haqqyi2wgncqa

Query Lattice for Translation Retrieval

Meiping Dong, Yong Cheng, Yang Liu, Jia Xu, Maosong Sun, Tatsuya Izuha, Jie Hao
2014 International Conference on Computational Linguistics  
Translation retrieval aims to find the most likely translation among a set of target-language strings for a given source-language string.  ...  Previous studies consider the single-best translation as a query for information retrieval, which may result in translation error propagation.  ...  f ) (2) where P (q|f ) is a translation sub-model, P (e|q, f ) is a retrieval sub-model, and Q(f ) is the set of all possible translations of the sentence f .  ... 
dblp:conf/coling/DongCLXSIH14 fatcat:37j4gvixwjhefpwqzf4lnznoui

Aspects of sentence retrieval

Vanessa Murdock
2007 SIGIR Forum  
Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently  ...  Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining  ...  We propose the use of translation models for the task of sentence retrieval for question answering.  ... 
doi:10.1145/1328964.1328990 fatcat:bilmicvjwvd3pnncu6tk6kpbry

Non-Parametric Adaptation for Neural Machine Translation [article]

Ankur Bapna, Orhan Firat
2019 arXiv   pre-print
We propose a novel n-gram level retrieval approach that relies on local phrase level similarities, allowing us to retrieve neighbors that are useful for translation even when overall sentence similarity  ...  We complement this with an expressive neural network, allowing our model to extract information from the noisy retrieved context.  ...  teams for their useful inputs and discussions.  ... 
arXiv:1903.00058v2 fatcat:3runmy6rbbfkhlh7437j3zsn4i

Non-Parametric Adaptation for Neural Machine Translation

Ankur Bapna, Orhan Firat
2019 Proceedings of the 2019 Conference of the North  
We propose a novel n-gram level retrieval approach that relies on local phrase level similarities, allowing us to retrieve neighbors that are useful for translation even when overall sentence similarity  ...  We complement this with an expressive neural network, allowing our model to extract information from the noisy retrieved context.  ...  teams for their useful inputs and discussions.  ... 
doi:10.18653/v1/n19-1191 dblp:conf/naacl/BapnaF19 fatcat:j52xwtup2ralpgg4b2vzrb5yyu

Retrieval-Based Transformer Pseudocode Generation

Anas Alokla, Walaa Gad, Waleed Nazih, Mustafa Aref, Abdel-Badeeh Salem
2022 Mathematics  
In this paper, a novel retrieval-based transformer pseudocode generation model is proposed.  ...  The proposed model adopts different retrieval similarity methods and neural machine translation to generate pseudocode.  ...  In [4] authors present a model for converting source code to pseudocode using Statistical Machine Translation (SMT).  ... 
doi:10.3390/math10040604 fatcat:izc6dqpiuzec5cn3p2jnqnchuq

Neural Machine Translation with Monolingual Translation Memory [article]

Deng Cai and Yan Wang and Huayang Li and Wai Lam and Lemao Liu
2021 arXiv   pre-print
First, the cross-lingual memory retriever allows abundant monolingual data to be TM. Second, the memory retriever and NMT model can be jointly optimized for the ultimate translation goal.  ...  memory retrieval in a cross-lingual manner.  ...  Retrieval Model The retrieval model is responsible for selecting the most relevant sentences for a source sentence from a large monolingual TM.  ... 
arXiv:2105.11269v2 fatcat:r6e5etqtbfaetijqxurpftcmwu

Search Engine Guided Non-Parametric Neural Machine Translation [article]

Jiatao Gu, Yong Wang, Kyunghyun Cho, Victor O.K. Li
2018 arXiv   pre-print
In the second stage--translation stage--, a novel translation model, called translation memory enhanced NMT (TM-NMT), seamlessly uses both the source sentence and a set of retrieved sentence pairs to perform  ...  In the first stage--retrieval stage--, an off-the-shelf, black-box search engine is used to retrieve a small subset of sentence pairs from a training set given a source sentence.  ...  For the proposed SEG-NMT model, one translation pair is retrieved from the training set.  ... 
arXiv:1705.07267v2 fatcat:4vmftrkpvzepfmzxd7mlrkoaqm

Chunk-based Nearest Neighbor Machine Translation [article]

Pedro Henrique Martins and Zita Marinho and André F. T. Martins
2022 arXiv   pre-print
However, kNN-MT requires retrieval for every single generated token, leading to a very low decoding speed (around 8 times slower than a parametric model).  ...  Semi-parametric models, which augment generation with retrieval, have led to impressive results in language modeling and machine translation, due to their ability to leverage information retrieved from  ...  For machine translation, Gu et al. (2018) introduced a semi-parametric model which uses an outof-the-box search engine to retrieve sentence-pairs with a similar source sentence, and incorporate these  ... 
arXiv:2205.12230v1 fatcat:g27gc3pmlbfzdcngrr3dcw7whm

WHU Question Answering System at NTCIR-8 ACLIA Task

Han Ren, Donghong Ji, Jing Wan
2010 NTCIR Conference on Evaluation of Information Access Technologies  
The system consists of a question translation model and a general question answering system for both factoid and complex questions.  ...  The translation model combines a translation engine and an online dictionary, which can provide more accurate translations of named entities in the questions.  ...  For this purpose, a translation model is implemented and EN-CS runs are submitted as well as CS-CS ones.  ... 
dblp:conf/ntcir/RenJW10 fatcat:gx66fm4oyfdind6qvx2q3hiroy

Improving Statistical Machine Translation Performance by Training Data Selection and Optimization

Yajuan Lü, Jin Huang, Qun Liu
2007 Conference on Empirical Methods in Natural Language Processing  
The online method adapts the translation model by redistributing the weight of each predefined submodels. Information retrieval model is used for the weighting scheme in both methods.  ...  Parallel corpus is an indispensable resource for translation model training in statistical machine translation (SMT).  ...  Online model weighting We also use TF-IDF information retrieval method for online model weighting. The procedure is as follows: For each input sentence: 1.  ... 
dblp:conf/emnlp/LuHL07 fatcat:glyvp5gj3rfnvboeclh65ujyse

Cross-lingual Retrieval for Iterative Self-Supervised Training [article]

Chau Tran, Yuqing Tang, Xian Li, Jiatao Gu
2020 arXiv   pre-print
We utilized these findings to develop a new approach -- cross-lingual retrieval for iterative self-supervised training (CRISS), where mining and training processes are applied iteratively, improving cross-lingual  ...  Using this method, we achieved state-of-the-art unsupervised machine translation results on 9 language directions with an average improvement of 2.4 BLEU, and on the Tatoeba sentence retrieval task in  ...  Acknowledgments and Disclosure of Funding We thank Angela Fan, Vishrav Chaudhary, Don Husa, Alexis Conneau, and Veselin Stoyanov for their valuable and constructive suggestions during the planning and  ... 
arXiv:2006.09526v2 fatcat:abklxi4sorfqdm357ofz2kjlnq
« Previous Showing results 1 — 15 out of 69,785 results