Nasiru Muhammad Dankolo

Improving Machine Translation Using an Efficient Dual-Bert Adversarial Network (DBAN) Model for User-Generated Content

An Efficient Dual-BERT Adversarial Network (DBAN) is proposed to improve the translation of noisy UGC by integrating contextual representation learning with adversarial training and significantly improves contextual understanding and cross-lingual semantic alignment while maintaining computational efficiency.