Smart Criminal Judgment
Analysis System
An advanced AI-driven research methodology developed for the Sri Lankan Penal Code, leveraging semantic legal extraction and predictive modeling architectures.
System Architecture
Resource Finder
Identify relevant laws from plain language case facts.
“Uses BERT-based semantic extraction to query the Penal Code and Criminal Procedure Code, clustering legal sections into Prosecution and Defence categories.”
Argument Builder
Generate structured case profiles from judgment documents.
“Performs deep document parsing to extract semantically similar past cases, building balanced 'Prosecution vs Defence' argument reports with strength scores.”
Outcome Predictor
Predict appeal results based on historical patterns.
“Employs Bayesian statistical modelling on thousands of historical judgments to calculate probability distributions for outcomes like Acquittal or Sentence Reduction.”
Semantic Chatbot
Multilingual research aid for real-time legal queries.
“An advanced NLP assistant that interprets and answers complex legal questions based on the insights generated by the system's core modules.”
The Research Problem
The Sri Lankan criminal justice system is currently slowed by a manual review bottleneck. Our methodology introduces advanced automation to process complex case facts and derive semantic legal insights.
Literature & Innovation
Bridging the gap between global LLMs and local SL litigation. We utilize Semantic Argument Clustering (SAC) as a primary methodology for the region.
Tech Stack
Project Milestones
Project Proposal
12%Initial concept presentation and project charter submission.
Progress Presentation 1
15%First technical milestone evaluating 50% functional completion.
Progress Presentation 2
18%Refinement phase assessing 90% system integration.
Demonstration
10%Final system walkthrough and poster presentation.
Final Assessment
10%Full dissertation submission and final defense.
Viva & Video
10%Final viva, commercialization video, and user testing results.
Resources & Archive
Our Team
Research Affiliation 2025
Connect
Open for technical collaborations and research inquiries regarding the SCJAS methodology.
Location
SLIIT, Malabe, Sri Lanka