Attention Based Bidirectional LSTM Model for Data-to-text Generation
Abhishek Kumar Pandey, Sanjiban Sekhar Roy · 2024
Natural Language Processing (NLP) helps process human language computationally. Recently, automatic data-to-text generation, article generation, has received much popularity. Deep learning has already revolutionized the applicability of natural language processing, and attention mechanism has changed how we work with them. This paper uses the attention mechanism for Bidirectional Long-Term Short-Term Memory (Bi-LSTM) by automatically taking input as words and generating a sentence. The generated text is validated with the BLEU Score. Moreover, Bi-LSTM with attention model has shown good BLEU scores compared to other models such as LSTM, LSTM with attention, LSTM, and Bi-LSTM. Results are 0.67, 0.42,0.21, and 0.24, respectively. The accuracy of Bi-LSTM with attention model is 92%