Litcius/Paper detail

English-Hindi Neural Machine Translation-LSTM Seq2Seq and ConvS2S

Gaurav Tiwari, Arushi Sharma, Aman Sahotra, Rajiv Kapoor

202054 citationsDOI

Abstract

Language Translation connects people from all around the world, to collaborate, share information and build relationships. Neural Machine Translation has been able to achieve significant improvement over primitive techniques: Rule-based and Statistical Machine Translation. This paper analyzes and compares two Neural Machine Translation models on the basis of different parameters for English-Hindi language pair: Sequence to Sequence Learning architecture with both encoder and decoder implemented using (1) Long Short Term Memory (LSTM) and (2) Convolutional Neural Network (CNN) with attention mechanism applied in both models. The analysis presents insights that reveal the models and parameters best for this task.

Topics & Concepts

Machine translationComputer scienceArtificial intelligenceNatural language processingConvolutional neural networkRecurrent neural networkTranslation (biology)HindiArtificial neural networkEncoderSequence (biology)Task (project management)Example-based machine translationSpeech recognitionMachine learningEngineeringGeneticsSystems engineeringMessenger RNAOperating systemChemistryBiochemistryGeneBiologyNatural Language Processing TechniquesTopic ModelingText and Document Classification Technologies
English-Hindi Neural Machine Translation-LSTM Seq2Seq and ConvS2S | Litcius