Abstract:
The judiciary is a strong pillar of Pakistan whose responsibility is to provide
justice. In recent years, Pakistani courts have digitized the legal system which
resulted in generating a large amount of data every day. As this data continues
to grow at a rapid rate, it has become essential to process this massive chunk
of data to better meet the requirements of the respective stakeholders. However,
extracting the required information from this unstructured legal text is the main
issue. Since Artificial Intelligence (AI) is finding its application in all domains of
our lives. The use of AI techniques can also be helpful in courtrooms. Therefore,
our focus is to build a machine learning based system to extract information from
semi-structured legal documents. Our system focuses on 15 entities including court
names, judge names, dates, case numbers, respondent names, reference cases,
person names, references, etc. Labeled datasets are acquired comprised of the
publicly available legal judgments from the Supreme Court of Pakistan and Lahore
High Court. In the first stage, entities are extracted and the relationship among
them is mapped. Later on, the deep learning model Neural Tensor Network (NTN)
is trained and fine-tuned on datasets to identify the entities in documents. Our
model has successfully outdone the previously published research.
