Efficient Word Alignment with Markov Chain Monte Carlo

Östling, Robert; Tiedemann, Jörg

doi:10.1515/pralin-2016-0013

Cited by 77 publications

(47 citation statements)

References 18 publications

(24 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We believe that one reason for the lack of consistent benefits of MTL in the labelling literature is that the proposed models share all or part of parameters for extracting hidden states, which leads to optimization conflicts when different tasks require different features. We believe it would be helpful if we make the model have ability to learn a task-specific representation (Ammar et al, 2016;Östling and Tiedemann, 2016;Kiperwasser and Ballesteros, 2018) at the same time. This observation led us to design a new LSTM cell which allows at almost no additional computation cost to efficiently train a single RNN-based model, where task-specific labelers clearly outperform their singly-tasked counterparts.…”

Section: Existing Shared Encoder Methodsmentioning

confidence: 99%

Untitled

Lu¹,

Bai

Langlais

2019

Proceedings of the 2019 Conference of the North

View full text Add to dashboard Cite

Multi-task learning (MTL) has been studied recently for sequence labeling. Typically, auxiliary tasks are selected specifically in order to improve the performance of a target task. Jointly learning multiple tasks in a way that benefit all of them simultaneously can increase the utility of MTL. In order to do so, we propose a new LSTM cell which contains both shared parameters that can learn from all tasks, and task-specific parameters that can learn task specific information. We name it a Shared-Cell Long-Short Term Memory (SC-LSTM). Experimental results on three sequence labeling benchmarks (named-entity recognition, text chunking, and part-of-speech tagging) demonstrate the effectiveness of our SC-LSTM cell.

show abstract

Section: Existing Shared Encoder Methodsmentioning

confidence: 99%

Untitled

Lu¹,

Bai

Langlais

2019

Proceedings of the 2019 Conference of the North

View full text Add to dashboard Cite

show abstract

“…First, we looked at the number of times the possessive pronoun son was translated by his or her in the training data. For this purpose, we used eflomal (Östling and Tiedemann, 2016) use the alignment link to find all possible translation of the French son token. 13 Results reported in Table 3 show that translations of son by his are three times more frequent than translations by her.…”

Section: Predicting Failurementioning

confidence: 99%

Screening Gender Transfer in Neural Machine Translation

Wisniewski¹,

Zhu²,

Bailler³

et al. 2021

Proceedings of the Fourth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP

View full text Add to dashboard Cite

This paper aims at identifying the information flow in state-of-the-art machine translation systems, taking as example the transfer of gender when translating from French into English.Using a controlled set of examples, we experiment several ways to investigate how gender information circulates in a encoder-decoder architecture considering both probing techniques as well as interventions on the internal representations used in the MT system. Our results show that gender information can be found in all token representations built by the encoder and the decoder and lead us to conclude that there are multiple pathways for gender transfer.

show abstract

“…As this can be achieved by CombAlign, it makes WAScore an effective mechanism for measuring parallelism. CombAlign uses the following tools in our experiment; (i) AWESoME (Dou and Neubig, 2021), (ii) eflomal (Östling and Tiedemann, 2016), and (iii) fast_align (Dyer et al, 2013). WAScore is calculated for each sentence using Equation (2):…”

Section: Sentence Scoringmentioning

confidence: 99%

Effective Bitext Extraction From Comparable Corpora Using a Combination of Three Different Approaches

Steingrímsson¹,

Lohar²,

Loftsson³

et al. 2021

Proceedings of the 14thWorkshop on Building and Using Comparable Corpora (BUCC 2021)

View full text Add to dashboard Cite

Parallel sentences extracted from comparable corpora can be useful to supplement parallel corpora when training machine translation (MT) systems. This is even more prominent in low-resource scenarios, where parallel corpora are scarce. In this paper, we present a system which uses three very different measures to identify and score parallel sentences from comparable corpora. We measure the accuracy of our methods in low-resource settings by comparing the results against manually curated test data for English-Icelandic, and by evaluating an MT system trained on the concatenation of the parallel data extracted by our approach and an existing data set. We show that the system is capable of extracting useful parallel sentences with high accuracy, and that the extracted pairs substantially increase translation quality of an MT system trained on the data, as measured by automatic evaluation metrics.

show abstract

Efficient Word Alignment with Markov Chain Monte Carlo

Cited by 77 publications

References 18 publications

Untitled

Untitled

Screening Gender Transfer in Neural Machine Translation

Effective Bitext Extraction From Comparable Corpora Using a Combination of Three Different Approaches

Contact Info

Product

Resources

About