With the power of Deep learning in NLP, I along with my team members exploited a powerful pre-trained transformer-based NLP model, GPT-2, to develop an application, TiGen, that can generate a list of suitable titles for research papers by taking only their abstracts as input. For training the said model, we curated our own dataset of title-abstract pairs. After successfully achieving high accuracy, we further dedicated ourselves to writing a research paper based on our findings which is yet to be published.

With the power of Deep learning in NLP, I along with my team members exploited a powerful pre-trained transformer-based NLP model, GPT-3, to develop a multi-lingual application, MUCE, that can generate a use cases and their corresponding actors by taking only the user's software requirements as input. For training the said model, we curated our own dataset of case studies collected from software engineering projects undertaken at the Department of Computer Science, SRCASW, DU. The application can also act as a feedback mechanism to improve and fine-tune software requirements to make them more complete. Our application works with great accuracy.