[META] Diffusion Boltzmann Sampler Implementation Tracker

[Sakeeb91]

Overview

This meta-issue tracks the implementation of Diffusion Models as Boltzmann Samplers - a neural approach to statistical mechanics sampling.

Project Summary

Train score-based diffusion models to sample from Boltzmann distributions, where the learned score function equals the physical force field. Targets: 2D Ising model and Lennard-Jones fluid with interactive React visualization.

Architecture

Frontend (React) ◄──► Backend (FastAPI) ◄──► ML Engine (PyTorch)
     │                     │                       │
  Plotly.js            WebSocket              Score Network
  Animations           REST API               Diffusion Sampler
  Controls             Streaming              MCMC Baseline

Implementation Phases

• Phase 0: Foundation & Setup [Phase 0] Foundation & Setup #2 ✅
• Phase 1: Ising Model Core [Phase 1] Ising Model Core #3 ✅
• Phase 2: Score Network Architecture [Phase 2] Score Network Architecture #4 ✅
• Phase 3: Training Loop [Phase 3] Training Loop #5 ✅
• Phase 4: Diffusion Sampling [Phase 4] Diffusion Sampling #6 ✅
• Phase 5: React Frontend Core [Phase 5] React Frontend Core #7 ✅
• Phase 6: Diffusion Animation [Phase 6] Diffusion Animation #8 ✅
• Phase 7: Analysis & Correlations [Phase 7] Analysis & Correlations #9 ✅
• Phase 8: Lennard-Jones Extension [Phase 8] Lennard-Jones Extension #10 ✅
• Phase 9: Polish & Documentation [Phase 9] Polish & Documentation #11

Key Deliverables

1. Ising Model Sampling - Neural sampler matching MCMC correlation functions
2. Interactive Visualization - Real-time diffusion animation, temperature controls
3. Lennard-Jones Extension - Continuous particle system with radial distribution function
4. Full-Stack Web App - FastAPI + React end-to-end

Success Criteria

• Neural sampler produces correct magnetization distribution P(M)
• Correlation functions match gold-standard MCMC within statistical error
• Interactive sampling in < 5 seconds for 32x32 lattice
• Diffusion animation renders smoothly at 20+ fps

Resources

• Implementation Plan
• README