Lexin’s research rests on a shared foundation (retrieval, reasoning, and trust and evaluation) and on infrastructure for Indonesian-language legal data. On top of it, four applied areas tackle real legal questions with partners.
“How can deeds and notarial services be created, verified and stored electronically with legal certainty?”
“How can rulings and labour regulation data help resolve disputes faster and more fairly?”
“What rules does Indonesia need for AI to be safe, fair and accountable?”
“What can machines learn from Indonesian court rulings, and how do we test it?”
Core research that supports all four areas. Every method is tested on Indonesian-language legal data before it is used in an applied area.
Finding the right articles, rulings and sources, and tying every answer back to its source.
Modelling how norms, facts and rulings relate, including legal entailment and argumentation.
Benchmarks, robustness tests and evaluation protocols so legal AI systems can be held accountable.
Reusable data and tooling for our team and the research community.
A catalogue of primary and secondary legal sources with licensing status and coverage.
Schemas and guidelines for annotating entities, relations and legal document structure.
Versioned corpora with clear test splits for fair experiments.
Code, configuration and results so others can repeat our experiments.
New research areas start with partners. If your institution has a legal problem that needs data and AI, let’s design it together.
