| ---
|
| title: MVM2 Math Verification System
|
| emoji: 🧮
|
| colorFrom: red
|
| colorTo: gray
|
| sdk: gradio
|
| sdk_version: 6.9.0
|
| app_file: app.py
|
| pinned: false
|
| license: mit
|
| short_description: MVM² High-Fidelity Glassmorphic Math Verification Dashboard
|
| ---
|
|
|
| # MVM² — Multi-Model Math Verification System
|
|
|
| A multi-agent AI system that verifies mathematical reasoning step-by-step using symbolic computation (SymPy) and LLM-based logic checking.
|
|
|
| 
|
| 
|
| 
|
| 
|
|
|
| ## Features
|
| - 🔬 **Parallel Verification**: Runs SymPy, LLM, and Ensemble models simultaneously
|
| - 📊 **Consensus Fusion**: Weighted scoring across multiple agents
|
| - 🧠 **Explainable AI**: Step-by-step error classification with natural language explanations
|
| - 📄 **Report Export**: Download verification reports as PDF, Word, or Markdown (powered by VibeDoc)
|
|
|
| ## How to Use
|
| 1. Enter your math problem and each step on a new line in the text box
|
| 2. Select which AI verifiers to enable in the sidebar
|
| 3. Click **Run Verification Pipeline**
|
| 4. Review the verdict, confidence score, and detailed breakdown
|
| 5. Download the full analysis report in your preferred format
|
|
|
| ## System Architecture
|
| ```
|
| Input → SymPy Symbolic Verifier ─┐
|
| → LLM Logical Checker ──→ Consensus Fusion → Verdict + Report
|
| → Ensemble Neural Check ─┘
|
| ```
|
|
|
| ## Based on Research Paper
|
| *Mathematical Reasoning Enhancement in Large Language Models* — VNRVJIET, Hyderabad |
