On April 4th 2023, at the International Hall of Hallym University, to discuss the Third-Year objectives and research status.
Hallym University is conducting research focused on data collection and generation, with a particular emphasis on collecting judgment documents and building datasets. They are currently working on collecting and structuring various crime-related data.
Additionally, the LIFS lab of Hallym University has collected 62,526 judgment cases related to various crimes such as murder, sexual assault, robbery, and fraud. The judgment case for murder and sexual assault have already been structured and shared through the NAS (Network Attached Storage), while the structuring work for robbery and fraud cases is still ongoing.
The research goal is to collect and structure an additional 20,000 cases each for sexual assault, robbery, and fraud this year. Constructing a dataset of similar sentences based on the crime judgment documents and currently engaged in dataset construction and labeling work. Also focus on constructing crime fact extraction datasets and collaborate with the police to generate such datasets. They utilize ChatGPT for data generation and verification and expect to have around 5,000 datasets in the future.
Lastly, they are conducting research on argument structure datasets. Having completed the analysis of 240 murder cases and 48 rape cases, the plan is to analyze robbery and theft cases next. They announced their intention to collaborate with KETI (Korea Electronics Technology Institute) to actively explore automation approaches for annotation using reference sets. They also mentioned their plans to define concepts for sub-classification of elements and relationships.”
Note: Some terms like “KETI” and “NAS” have been left as they are since they are acronyms or specific entities that might not have direct translations.