International Conference on Information and Knowledge Technology

فارسی

Home / شانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش

Integrating Wasserstein GANs for High-Speed Transformer-Based Neural Machine Translation

Authors :

Parisa Nekoogol¹ Mostafa Salehi²

1- دانشگاه تهران 2- دانشگاه تهران

Keywords :

Neural Machine Translation،Generative Adversarial Networks،Reinforcement Learning،Transformer

Abstract :

Neural machine translation (NMT), a key achievement in natural language processing (NLP), continues to face challenges such as producing low-quality output for complex sentences and lacking natural fluency. This study aimed to improve machine translation quality by integrating Generative Adversarial Networks (GANs) with an NMT model. Initially, the baseline NMT model, derived from previous research and based on recurrent neural networks (RNNs), was reconstructed and implemented. Subsequently, this architecture was replaced with the advanced Transformer architecture, and the system was developed using a Wasserstein Generative Adversarial Network (WGAN). To overcome the crucial problem of textual data discontinuity (non-differentiability), the Self-Critical Sequence Training (SCST) method, a reinforcement learning (RL) algorithm, was employed. A core objective was to analyze the performance benefits of adversarial training when applied to a robust Transformer-based generator. The research concluded that while adversarial training enhances the model's performance in generating more fluent translations, this particular improvement is more substantial and notable for models based on recurrent neural networks compared to the Transformer architecture.

List of archived papers

Silicon photonic microring resonators: A Novel optical router based on Negative-First routing algorithm

Negin Bagheri Renani - Elham Yaghoubi

پیش‌بینی فضایی–زمانی و مقایسه ریسک تب دنگی با استفاده از مدل‌های یادگیری عمیق LSTM و GRU و مدل یادگیری ماشین Random Forest بر پایه مؤلفه‌های اقلیمی و مکانی

محمد بابائی - نجمه نیسانی سامانی

طبقه بندی روش های شناسایی داده های تکراری در جهت تسهیل فرایند پاکسازی داده ها

مهدی جعفری - احمد عبدالله زاده بار فروش

تحلیل سازه‌های موثر بر پذیرش فناوری بلاکچین و استفاده از آن در صنعت بیمه ایران با استفاده از تکنیک معادلات ساختاری (مطالعه موردی: شرکت کارگزاری رسمی بیمه زندگی خوب)

احسان هنری - آفرین اخوان

Robustness Gap in NLP Models for Vulnerability Descriptions: Benchmarking and Data Augmentation

AmirHossein Majd - Mahdi Yousefikia - Saghar Ghasemzadeh - Amirreza Asari - Arya Khoshnavataher - Seyedeh Leili Mirtaheri

A High-Speed Quantum Reversible Controlled Adder/Subtractor Circuit

Negin Mashayekhi - Mohammad Reza Reshadinezhad - Shekoofeh Moghimi

An Adaptive Mutation-Enhanced EHO-SVM Framework for Intrusion Detection in IoMT Environments

Amirhossein Damia - Erfaneh Khanmohammadi

A hybrid CNN–transformer framework for retinal disease classification

Hanie Zomorrodi - Hassan Khotanlou

پیش‌بینی میزان بقای بیماران مبتلا به سرطان ریه با استفاده از ترکیب کارآمد روش‌های داده‌کاوی و بهینه‌سازی رقابت استعماری

رخشان رمضانی سرچشمه - مهدی هاشم‌زاده - امین گلزاری اسکوئی

Particle Swarm Optimization-Based Framework for 3D Swarm Robotic Navigation Using Artificial Potential Field Dynamics

Samim Kamyab - Masoud Shirzadeh - Ghoncheh Zand

more

Samin Hamayesh - Version 42.5.2