WILO-BiLSTM: Optimization Based Bidirectional Long Short-Term Memory Model for Intelligent Virtual Assistant System
TL;DR
The proposed WILO-BiLSTM model can perform superior to the conventional approaches and its performance results in terms of METEOR, BLEU, ROUGE, and SPICE score at training data 90% is 0.28, 0.50, 0.56, and 26.93 for the SquAD dataset, respectively.
Abstract
In the field of Natural Language Processing (NLP), intelligent virtual assistance systems or information retrieval systems have earned significant attention in recent years as they assist users and solve their queries by providing accurate answers. However, numerous studies were developed for intelligent virtual assistance systems, but they face limitations like multilingual aspects, ambiguity resolution, lack of consistency in answers, complex language structures and linguistic variations. Hence, to tackle these constraints, this research proposed an intelligent virtual assistant system using Wolf Interactive Learning Optimization-based Bidirectional Long Short-Term Memory (WILO-BiLSTM) model. The integration of BiLSTM in the model enhances the understanding of complex questions and captures semantic information, resulting in improved model performance. In addition, it processes sequential data effectively and overcomes gradient problems in long sequences. Moreover, the exploitation of the WILO algorithm in the model optimizes BiLSTM hyperparameters and significantly enhances the convergence speed with reduced computational cost. Notably, the experimental outcomes reveal that the proposed WILO-BiLSTM model can perform superior to the conventional approaches and its performance results in terms of METEOR, BLEU, ROUGE, and SPICE score at training data 90% is 0.28, 0.50, 0.56, and 26.93 for the SquAD dataset, respectively.