ARD (Autonomous Research Discovery)

ARD is a Python package for building, curating, and mining knowledge graphs to enable autonomous research discovery. It's part of the BeeARD ecosystem that aims to accelerate scientific progress through AI-driven hypothesis generation and validation.

🚀 Features

• Knowledge Graph Management: Build and maintain comprehensive knowledge graphs from scientific literature
• Multi-Agent Systems: Leverage both AutoGen and LangGraph implementations for hypothesis generation
• Subgraph Mining: Extract and analyze meaningful subgraphs for research insights
• Hypothesis Generation: Generate novel research hypotheses using advanced LLM-based agents
• CLI Interface: Command-line tools for common knowledge graph operations

📦 Installation

ARD requires Python 3.12+ and uses UV as its package manager.

# Clone the repository
git clone https://github.com/ARDSys/hackathon.git
cd hackathon

# Setup virtual env
uv sync
source .venv/bin/activate

# Install in development mode
uv pip install -e .

🛠️ Usage

Building Knowledge Graphs

from ard.knowledge_graph import KnowledgeGraph
from ard.data import DatasetItem

# Initialize a knowledge graph
kg = KnowledgeGraph()

# Add data and build relationships
# ... (see examples/ for detailed usage)

Generating Hypotheses

ARD provides two workflow implementations for hypothesis generation:

1. AutoGen-based Workflow:

from hackathon.autogen import generate_hypothesis

# Generate hypotheses using AutoGen agents
hypothesis = generate_hypothesis.run(subgraph, output_dir="results")
# Access hypothesis properties: hypothesis.title, hypothesis.statement, hypothesis.references, etc.

2. LangGraph-based Workflow:

from hackathon.langgraph import generate_hypothesis

# Generate hypotheses using LangGraph agents
hypothesis = generate_hypothesis.run(subgraph, output_dir="results")
# Access hypothesis properties: hypothesis.title, hypothesis.statement, hypothesis.references, etc.

CLI Usage

# Create a knowledge graph
ard graph --data-path /path/to/data --output knowledge_graph.pkl

# Extract subgraphs
ard subgraph --graph-path knowledge_graph.pkl --output-dir output

🏗️ Architecture

ARD is organized into several key components:

• knowledge_graph/: Core knowledge graph implementation and management
• subgraph/: Subgraph extraction and analysis tools
• hypothesis/: Hypothesis generation and validation
• data/: Data ingestion and management
• llm/: LLM integration utilities
• utils/: Common utilities and helpers
• storage/: Storage backends and persistence

📚 Documentation

For detailed documentation, visit docs.beeard.ai.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.