The goal of Part I of the task is to use raw textual data in language models for recommendation$-$based application.