# A universal vision transformer for fast calorimeter simulations

arXiv: https://arxiv.org/abs/2601.05289 · HTML: https://mlsanity.com/paper/2601.05289/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 37 references that are in the map.
- First posted on arXiv 7 Jan 2026.

## Builds on (most cited first, 7 of 7)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains](https://mlsanity.com/paper/2006.10739/) (2020)
- [CaloMan: Fast generation of calorimeter showers with density estimation on learned manifolds](https://mlsanity.com/paper/2211.15380/) (2022)
- [Bumblebee: Foundation Model for Particle Physics Discovery](https://mlsanity.com/paper/2412.07867/) (2024)
- [Bespoke Non-Stationary Solvers for Fast Sampling of Diffusion and Flow Models](https://mlsanity.com/paper/2403.01329/) (2024)
- [Scaling Up Diffusion and Flow-based XGBoost Models](https://mlsanity.com/paper/2408.16046/) (2024)

## Built on by (most cited first, 2 of 2)

- [AllShowers: One model for all calorimeter showers](https://mlsanity.com/paper/2601.11716/) (2026)
- [CaloTrilogy: Toward a Breakthrough in One-Step, End-to-End, Physics-Guided Shower Generation for Modern Calorimeters](https://mlsanity.com/paper/2606.04165/) (2026)

## Abstract (arXiv)

The high-dimensional complex nature of detectors makes fast calorimeter simulations a prime application for modern generative machine learning. Vision transformers (ViTs) can emulate the Geant4 response with unmatched accuracy and are not limited to regular geometries. Starting from the CaloDREAM architecture, we demonstrate the robustness and scalability of ViTs on regular and irregular geometries, and multiple detectors. Our results show that ViTs generate electromagnetic and hadronic showers with minimal deviations from Geant4 in multiple evaluation metrics, while maintaining the generation time in the O(10-100) ms on a single GPU. Furthermore, we show that pretraining on a large dataset and fine-tuning on the target geometry leads to reduced training costs and higher data efficiency, or altogether improves the fidelity of generated showers.