Assessing work from various fields is necessary to analyze & survey an institution's performance over a certain period of time having progression in multiple divisions. Many necessary characteristics such as Impact Factor (IF) of acclaimed works are influenced by non-uniform distribution of publications in different sections and renowned journals. Classifying file elements with NLP based on the publication titles would be supportive and intuitive. Text Analysis and Field Classification requires a large amount of data for a model to be trained and efficient. So, a Zero shot learning approach is opted to distinguish various publications into their respective sectors. Unlike other models, this model is enhanced to leverage CSV format files for both input and output. Different Pre-Trained Language models have been used and their performances are recorded. The advantage of zero shot learning over regular methods is discussed.

Natural Language Processing Pre-Trained Language Model Text Classification Transformers Zero Shot Learning