Generative Models for Fast Simulation of Cherenkov Detectors at the Electron-Ion Collider

2025-04-26

AI Summary

Traditional Geant4-based optical photon transport simulation for Cherenkov detectors—e.g., the hpDIRC at the Electron–Ion Collider (EIC)—incurs prohibitive computational cost, hindering high-precision, large-scale particle identification (PID). To address this, we propose the first open-source generative fast-simulation framework tailored for Cherenkov detectors. Our method integrates conditional GANs with diffusion models, geometry-aware light-propagation modeling, CUDA-accelerated response synthesis, and physics-informed loss functions. The resulting framework enables GPU-accelerated, geometry-agnostic, and scalable high-fidelity simulation. It achieves >1000× speedup over Geant4 while maintaining PID-critical feature errors below 2%, and supports unlimited sample generation across the full EIC acceptance. The implementation is publicly released, establishing a reproducible, collaborative simulation infrastructure for AI-driven particle identification.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Web Mining and Content Analysis: Web data generation and simulation
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

The integration of Deep Learning (DL) into experimental nuclear and particle physics has driven significant progress in simulation and reconstruction workflows. However, traditional simulation frameworks such as Geant4 remain computationally intensive, especially for Cherenkov detectors, where simulating optical photon transport through complex geometries and reflective surfaces introduces a major bottleneck. To address this, we present an open, standalone fast simulation tool for Detection of Internally Reflected Cherenkov Light (DIRC) detectors, with a focus on the High-Performance DIRC (hpDIRC) at the future Electron-Ion Collider (EIC). Our framework incorporates a suite of generative models tailored to accelerate particle identification (PID) tasks by offering a scalable, GPU-accelerated alternative to full Geant4-based simulations. Designed with accessibility in mind, our simulation package enables both DL researchers and physicists to efficiently generate high-fidelity large-scale datasets on demand, without relying on complex traditional simulation stacks. This flexibility supports the development and benchmarking of novel DL-driven PID methods. Moreover, this fast simulation pipeline represents a critical step toward enabling EIC-wide PID strategies that depend on virtually unlimited simulated samples, spanning the full acceptance of the hpDIRC.

Problem

• Accelerating Cherenkov detector simulation using generative models
• Reducing computational cost of optical photon transport simulation
• Enabling scalable particle identification with GPU-accelerated fast simulation

Innovation

• Generative models accelerate Cherenkov detector simulations
• GPU-accelerated alternative to Geant4 for particle identification
• Open standalone tool for high-fidelity on-demand datasets