# A reconfigurable neural network ASIC for detector front-end data compression at the HL-LHC

arXiv: https://arxiv.org/abs/2105.01683 · HTML: https://mlsanity.com/paper/2105.01683/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- Builds on 2 of its 31 references that are in the map.
- First posted on arXiv 4 May 2021.

## Builds on (most cited first, 2 of 2)

- [Fast inference of deep neural networks in FPGAs for particle physics](https://mlsanity.com/paper/1804.06913/) (2018)
- [Machine Learning in High Energy Physics Community White Paper](https://mlsanity.com/paper/1807.02876/) (2018)

## Built on by (most cited first, 20 of 20)

- [Lorentz group equivariant autoencoders](https://mlsanity.com/paper/2212.07347/) (2022)
- [Ultra-low latency recurrent neural network inference on FPGAs for physics applications with hls4ml](https://mlsanity.com/paper/2207.00559/) (2022)
- [FastML Science Benchmarks: Accelerating Real-Time Scientific Edge Machine Learning](https://mlsanity.com/paper/2207.07958/) (2022)
- [Snowmass 2021 Computational Frontier CompF03 Topical Group Report: Machine Learning](https://mlsanity.com/paper/2209.07559/) (2022)
- [On-Sensor Data Filtering using Neuromorphic Computing for High Energy Physics Experiments](https://mlsanity.com/paper/2307.11242/) (2023)
- [Neural approximation of Wasserstein distance via a universal architecture for symmetric and factorwise group invariant functions](https://mlsanity.com/paper/2308.00273/) (2023)
- [Machine-Learning Compression for Particle Physics Discoveries](https://mlsanity.com/paper/2210.11489/) (2022)
- [Physics Community Needs, Tools, and Resources for Machine Learning](https://mlsanity.com/paper/2203.16255/) (2022)
- [Architectural Implications of Neural Network Inference for High Data-Rate, Low-Latency Scientific Applications](https://mlsanity.com/paper/2403.08980/) (2024)
- [Reliable edge machine learning hardware for scientific applications](https://mlsanity.com/paper/2406.19522/) (2024)
- [CGRA4ML: A Hardware/Software Framework to Implement Neural Networks for Scientific Edge Computing](https://mlsanity.com/paper/2408.15561/) (2024)
- [Hardware-Aware Tensor Networks for Real-Time Quantum-Inspired Anomaly Detection at Particle Colliders](https://mlsanity.com/paper/2603.26604/) (2026)
- [PulseDL-II: A System-on-Chip Neural Network Accelerator for Timing and Energy Extraction of Nuclear Detector Signals](https://mlsanity.com/paper/2209.00884/) (2022)
- [Differentiable Earth Mover's Distance for Data Compression at the High-Luminosity LHC](https://mlsanity.com/paper/2306.04712/) (2023)
- [Embedded FPGA Developments in 130nm and 28nm CMOS for Machine Learning in Particle Detector Readout](https://mlsanity.com/paper/2404.17701/) (2024)
- [Data-driven detector signal characterization with constrained bottleneck autoencoders](https://mlsanity.com/paper/2203.04604/) (2022)
- [Investigating Resource-efficient Neutron/Gamma Classification ML Models Targeting eFPGAs](https://mlsanity.com/paper/2404.14436/) (2024)
- [Automated and Holistic Co-design of Neural Networks and ASICs for Enabling In-Pixel Intelligence](https://mlsanity.com/paper/2407.14560/) (2024)
- [Variable Rate Neural Compression for Sparse Detector Data](https://mlsanity.com/paper/2411.11942/) (2024)
- [Design Rules for Extreme-Edge Scientific Computing on AI Engines](https://mlsanity.com/paper/2604.19106/) (2026)

## Abstract (arXiv)

Despite advances in the programmable logic capabilities of modern trigger systems, a significant bottleneck remains in the amount of data to be transported from the detector to off-detector logic where trigger decisions are made. We demonstrate that a neural network autoencoder model can be implemented in a radiation tolerant ASIC to perform lossy data compression alleviating the data transmission problem while preserving critical information of the detector energy profile. For our application, we consider the high-granularity calorimeter from the CMS experiment at the CERN Large Hadron Collider. The advantage of the machine learning approach is in the flexibility and configurability of the algorithm. By changing the neural network weights, a unique data compression algorithm can be deployed for each sensor in different detector regions, and changing detector or collider conditions. To meet area, performance, and power constraints, we perform a quantization-aware training to create an optimized neural network hardware implementation. The design is achieved through the use of high-level synthesis tools and the hls4ml framework, and was processed through synthesis and physical layout flows based on a LP CMOS 65 nm technology node. The flow anticipates 200 Mrad of ionizing radiation to select gates, and reports a total area of 3.6 mm^2 and consumes 95 mW of power. The simulated energy consumption per inference is 2.4 nJ. This is the first radiation tolerant on-detector ASIC implementation of a neural network that has been designed for particle physics applications.