tharindudr/thesis

Deep Learning based Semantic Textual Similarity Metric for Applications in Translation Technology

Deep Learning based Semantic Textual Similarity Metric for Applications in Translation Technology

This repository contains the LaTeX source for the PhD thesis written by Tharindu Ranasinghe. Code for running the experiments presented in the thesis can be found in each respective chapter. Parts of this thesis appeared in the following peer-reviewed papers. For many of the topics discussed in both, the thesis provides more extensive coverage with more datasets. Code for running the experiments presented in these publications can be found in each respective publication's repository, linked below.

Publications

An Exploratory Analysis of Multilingual Word-Level Quality Estimation with Cross-Lingual Transformers Tharindu Ranasinghe, Constantin Orasan, and Ruslan Mitkov ACL 2021

TransQuest: Translation Quality Estimation with Cross-lingual Transformers Tharindu Ranasinghe, Constantin Orasan, and Ruslan Mitkov COLING 2020

TransQuest at WMT2020: Sentence-Level Direct Assessment Tharindu Ranasinghe, Constantin Orasan, and Ruslan Mitkov WMT 2020

Intelligent Translation Memory Matching and Retrieval with Sentence Encoders Tharindu Ranasinghe, Constantin Orasan, and Ruslan Mitkov EAMT 2020

Semantic textual similarity based on deep learning: Can it improve matching and retrieval for Translation Memory tools? Tharindu Ranasinghe, Ruslan Mitkov, Constantin Orasan, and Rocío Caro Quintana Corpora in Translation and Contrastive Research in the Digital Age: Recent advances and explorations John Benjamins

Semantic Textual Similarity with Siamese Neural Networks Tharindu Ranasinghe, Constantin Orasan, and Ruslan Mitkov RANLP 2019

Enhancing Unsupervised Sentence Similarity Methods with Deep Contextualised Word Representations Tharindu Ranasinghe, Constantin Orasan, and Ruslan Mitkov RANLP 2019