Browsing by Author "Khalaf, Ayesheh Ahrari"

One of the objectives and aspirations of scientists and engineers ever since the development of computers has been to interact naturally with machines. Hence features of artificial intelligence (AI) like natural language processing and natural language generation were developed. The field of AI that is thought to be expanding the fastest is interactive conversational systems. Numerous businesses have created various Virtual Personal Assistants (VPAs) using these technologies, including Apple's Siri, Amazon's Alexa, and Google Assistant, among others. While an ongoing effort to increase the friendliness and constancy of informal dialogue systems, most research focuses solely on simulating human-like replies, leaving the features of modeling interlocutors' awareness are unexplored. Meanwhile, cognitive science research reveals that awareness is a crucial indicator of a high-quality informal conversation. To precisely model understanding, Persona Perception (P2) Bot was developed using a transmitter-receiver-based structure. P2 Bot leverages mutual persona perception to improve the quality of customized dialogue generation. Even though many chatbots have been introduced through the years to diagnose or treat psychological disorders, we are yet to have a user-friendly chatbot available. This research aims on improving the quality of conversation generation by implementing the Generative Pre-trained Transformer-2 (GPT-2) model on P2 Bot. GPT-2 is a 1.5B parameter transformer model which produces state-of-the-art accuracy in a zero-shot setting on 7 out of 8 evaluated language modeling datasets. Observations on a large open-source dataset, PERSONA-CHAT, show that the technique is successful, with some improvement above state-of-the-art baselines in both automatic measures and human assessments. The model has achieved 82.2% accuracy on Hits@1(%) in the original data and 68.8% on the revised data. On the human evaluation, the model scored an average of 2.66 meaning the provided responses were coherent and informative. A smart generative cognitive behavioral therapy with spoken dialogue systems support was then developed using the model, which was then implemented using modern technologies in VPAs like voice recognition, Natural Language Understanding (NLU), and text-to-speech. This system is a magnificent device to help with voice-based systems because it can have therapeutic discussions with the users utilizing text and vocal interactive user experience.