Interview Q&A System using Natural Language Processing
Disciplines
Other Computer Sciences
Abstract (300 words maximum)
Recent advancements in Question-Answering (QA) systems have exploited natural language processing models to create human-like machine interaction across diverse domains. This paper delves into the persistent challenges of constructing domain-specific QA systems, with a keen focus on the one application: the job interview. We introduce a refined model tailored to grasp the nuances of our interview-centric dataset. We focus on the improvement of existing models—Distilled Bidirectional Encoder Representations from Transformers (DistilBERT) and Bidirectional Encoder Representations from Transformers (BERT)—and our objective is to tailor these architectures to the specificities of interview-based QA queries. Through rigorous evaluation metrics, we will present the merits of our optimized model, underscoring its potential in revolutionizing interview preparation methodologies.
Academic department under which the project should be listed
CCSE - Computer Science
Primary Investigator (PI) Name
Md Abdullah Al Hafiz Khan
Additional Faculty
Md Abdullah Al Hafiz Khan, Computer Science, mkhan74@kennesaw.view.usg.edu
Interview Q&A System using Natural Language Processing
Recent advancements in Question-Answering (QA) systems have exploited natural language processing models to create human-like machine interaction across diverse domains. This paper delves into the persistent challenges of constructing domain-specific QA systems, with a keen focus on the one application: the job interview. We introduce a refined model tailored to grasp the nuances of our interview-centric dataset. We focus on the improvement of existing models—Distilled Bidirectional Encoder Representations from Transformers (DistilBERT) and Bidirectional Encoder Representations from Transformers (BERT)—and our objective is to tailor these architectures to the specificities of interview-based QA queries. Through rigorous evaluation metrics, we will present the merits of our optimized model, underscoring its potential in revolutionizing interview preparation methodologies.