OCW Contextual ASR is a lecture-based dataset for contextual automatic speech recognition research in the education domain. It is constructed from two MIT OpenCourseWare courses and is intended to support the evaluation of ASR with external contextual information such as lecture slides and named entities. The dataset is introduced in the following paper: Improving Named Entity Transcription with Contextual LLM-based Revision（… See the full description on the dataset page: https://huggingface.co/datasets/lucille0he/ocw.
