A Trusted Human-Multi-Agent Reinforcement Learning Interaction Framework
This repository implements a Multi-Agent System (MAS) framework for human-machine collaborative crisis response, combining vision-language models (VL) and reinforcement learning (RL) to enhance safety and reliability. The framework features:
- Real-Time Task Execution: Modular task chains with built-in safety rules and human oversight.
- Simulation Training: Experience replay library for risk prediction and optimization.
- Dynamic Trust Mechanism: Balances task utility and safety constraints through RL.
Key Contributions:
- Dual-mode architecture (online execution + offline simulation).
- First fine-tuned safe LLM and training dataset for emergency scenarios.
- 15% improvement in helpfulness and 40% reduction in risk response rate compared to baseline.
git clone https://github.com/erwinmsmith/SOMAS.git
cd SOMAS
pip install -r requirements.txt- Real-Time Task Execution:
main.py --online- Simulation Training:
main.py --offline-
Online Execution System
- Planning-Execution Pipeline: Modular task chains drive tool operations.
- Safety Guardrails: Predefined rules and GPT-4-based risk assessment.
-
Offline Simulation System
- Task Generation: Synthetic tasks from manual records and prior knowledge.
- Experience Replay: Optimizes RL policies for dynamic environments.
| Domain | Model | Safety (↑) | Helpfulness (↑) | Risk Response Rate (↓) |
|---|---|---|---|---|
| Safety-CV | Qwen2-7B-VL | 4.5 | 4.7 | 40% |
- VL models reduced operational risks by 30% via image semantic parsing.
- Dynamic safety validation improved helpfulness by 15% over ToolEmu.
If you need a detailed data for Safty(train or test), contact me duanzhenke@sscapewh.com