About the dataset

The dataset comprises recordings of several real-world criminal trials as well as multiple civil televised litigations appeared on TV show "Judge Judy".


We manually annotated the transcripts of five real-world trials and 19 cases on "Judge Judy" TV show. For each transcript we extracted a set of propositions made by witnessess and other participants in the case and produced a set of logical constraints from those propositions. The JUST dataset contains propositions, constraints, as well as raw audio data from the said trials.

Download the dataset

Pre-publication preview dataset can be downloaded here (374 MB).