Schema Miner - Initial Schema Mining

Welcome to Schema Miner — a Human-in-the-Loop framework for Scientific Schema Mining using Large Language Models (LLMs). This tool is designed to assist you in extracting structured schema representations for scientific processes by leveraging scientific literature and continuous human feedback. This interactive demo will guide you step-by-step, starting from unstructured textual information and culminating in a well-defined, structured schema. Each tab above corresponds to a distinct stage in the Schema Miner workflow, progressively refining and enriching the extracted schema with domain knowledge. In this first stage, the LLM generates an initial JSON schema that captures essential properties, data types, and associated constraints relevant to the target scientific domain. This foundational schema acts as the basis for subsequent stages, where it is iteratively improved using expert input and ontological alignment. To get started, simply enter your OpenAI API key, provide a brief description of the target process, and upload a process specification document. Once submitted, the system will automatically generate an initial schema representation, marking the beginning of your schema discovery journey.

Hello! I am Schema Miner, your assistant for extracting an initial schema from a scientific process specification. To begin, could you please provide the name of the scientific process you are working with?

Extracted JSON Schema

Below is the initial schema representation automatically generated from your Process Specification Document. Please review the extracted schema carefully and proceed to the next stage for further refinement and validation with domain-specific knowledge.