1.3. Contributions

Through the making of this work, three separate types of assets were produced, and promptly delivered available to appropriate recipients, under the form of contributions. Firstly, we developed multiple datasets from the multiple Portuguese legal documents available and published them on the Internet.

These datasets are publicly available in the HuggingFace platform, allowing them to be easily utilized in other scenarios. Secondly, numerous language models adapted to the Portuguese legal domain were developed and made available on the same platform. Finally, we developed a semantic search system with the intent of helping judges in the court decision process.