The court clerical staff currently take several days to draft a comprehensive court order even after they have collected all relevant information from all the required parties. Most judgements follow a predefined structure including summarizing/structuring all content (this doesn't require any higher reasoning usually and is very repetitive work). The chatbot aims to automate this effort.
Any required content could be sourced through a form or through a conversation that fills in any missing information.
The bot then would generate a judgement using the content in a similar style as the previous judgements and without stating any falsities or creating its own conclusions. This should ideally be completed in less than 10 minutes.
Goals & Mid-Point Milestone
[ ] Setup and Installation
[ ] Complete setup and installation of chatbot framework and dependencies.
[ ] Conversational Agent Development
[ ] Develop a functional conversational agent for data gathering.
[ ] Document Structure Definition
[ ] Establish a structured document format for judgments.
[ ] LLM Training and Fine-Tuning
[ ] Train LLM using relevant datasets.
[ ] Fine-tune LLM to match the style of previous judgments.
[ ] Document Drafting Automation
[ ] Automate drafting process for generating coherent judgments.
[ ] Performance Evaluation
[ ] Benchmark LLM performance to ensure validity and efficiency.
Setup/Installation
No response
Expected Outcome
The anticipated result is a finely-tuned Language Model (LLM) capable of generating documents in the style of previous judgments. This LLM, integrated with a conversational agent, will streamline data gathering and document drafting processes, achieving completion within a targeted 10-minute timeframe. Through benchmarking on tasks like judgment validity, the LLM's reliability and accuracy will be validated, ultimately enhancing efficiency in government office operations.
Acceptance Criteria
No response
Implementation Details
This initiative will encompass the following key steps:
Defining Document Structure: Establishing a generic document structure and creating a comprehensive registry encompassing document types, forms, styles, and other relevant attributes.
Building a Conversational Agent: Developing a sophisticated conversational agent capable of seamlessly gathering pertinent information required for document drafting purposes.
Document Drafting: Automating the document drafting process to efficiently assemble collected data into coherent and comprehensive documents, adhering to predefined structures and formats.
Fine-Tuning LLM: Precision-tuning a Language Model (LLM) to closely align with the style and conventions observed in previous judgments, ensuring consistency and professionalism in generated documents.
Dataset Generation: Generating datasets derived from older documents to facilitate effective training and refinement of the LLM, enriching its understanding of language nuances and legal terminology.
Benchmarking LLM Performance: Evaluating the performance of the LLM through rigorous benchmarking exercises, particularly on complex tasks such as judgment validity, to validate its reliability and efficacy in document generation tasks.
Ticket Contents
The court clerical staff currently take several days to draft a comprehensive court order even after they have collected all relevant information from all the required parties. Most judgements follow a predefined structure including summarizing/structuring all content (this doesn't require any higher reasoning usually and is very repetitive work). The chatbot aims to automate this effort.
Any required content could be sourced through a form or through a conversation that fills in any missing information. The bot then would generate a judgement using the content in a similar style as the previous judgements and without stating any falsities or creating its own conclusions. This should ideally be completed in less than 10 minutes.
Goals & Mid-Point Milestone
[ ] Setup and Installation
[ ] Conversational Agent Development
[ ] Document Structure Definition
[ ] LLM Training and Fine-Tuning
[ ] Document Drafting Automation
[ ] Performance Evaluation
Setup/Installation
No response
Expected Outcome
The anticipated result is a finely-tuned Language Model (LLM) capable of generating documents in the style of previous judgments. This LLM, integrated with a conversational agent, will streamline data gathering and document drafting processes, achieving completion within a targeted 10-minute timeframe. Through benchmarking on tasks like judgment validity, the LLM's reliability and accuracy will be validated, ultimately enhancing efficiency in government office operations.
Acceptance Criteria
No response
Implementation Details
This initiative will encompass the following key steps:
Defining Document Structure: Establishing a generic document structure and creating a comprehensive registry encompassing document types, forms, styles, and other relevant attributes.
Building a Conversational Agent: Developing a sophisticated conversational agent capable of seamlessly gathering pertinent information required for document drafting purposes.
Document Drafting: Automating the document drafting process to efficiently assemble collected data into coherent and comprehensive documents, adhering to predefined structures and formats.
Fine-Tuning LLM: Precision-tuning a Language Model (LLM) to closely align with the style and conventions observed in previous judgments, ensuring consistency and professionalism in generated documents.
Dataset Generation: Generating datasets derived from older documents to facilitate effective training and refinement of the LLM, enriching its understanding of language nuances and legal terminology.
Benchmarking LLM Performance: Evaluating the performance of the LLM through rigorous benchmarking exercises, particularly on complex tasks such as judgment validity, to validate its reliability and efficacy in document generation tasks.
Mockups/Wireframes
No response
Product Name
Some Name
Organisation Name
Bandhu
Domain
Financial Inclusion
Tech Skills Needed
Flask, Python
Mentor(s)
some mentor
Category
Data Science