Industrial foundation models enhance real-time particle detector data processing
Leveraging Industrial Foundation Models at the Edge of Particle Physics Detectors via Distillation Learning and Hardware Co-design
Machine Learning
Summary
Big AI models can help handle the huge data from particle physics experiments better, but they are often too slow or large to run where data is collected. This work shows how to fine-tune a large time-series model from Google to do fast and accurate data analysis directly on detectors. The authors then shrink this model and design it to run efficiently on special hardware called FPGAs. Their smaller model works well in real time and can be used for many waveform analysis tasks beyond particle physics.
What this means in practice
- •For particle physics experiment teams: Run efficient AI-based waveform analysis directly on detector hardware for improved real-time data filtering and feature extraction.
- •For industrial engineers for time series data: Compress and deploy foundation models for 1D waveform regression tasks in real-time embedded systems using FPGA implementations.
Authors
Gia Ancone, Qibin Liu, Liangyu Wu, Julia Gonski
Abstract
Data acquisition (DAQ) systems at future particle physics experiments stand to benefit from the extremes of AI/ML development: large-scale foundation models can enhance the performance of feature extraction algorithms, and small-scale on-detector deployments can enable real-time intelligent data handling. This work provides the first fine-tuning of an industrial foundation model for particle physics DAQ. Starting from the backbone of Google Research's TimesFM (Time Series Foundation Model), we demonstrate fine-tuning on real-time regression tasks for drift chamber trackers and dual-readout calorimeters. Furthermore, the fine-tuned TimesFM model is distilled into a student and co-designed with FPGA implementation to enable these models to run in real-time at future colliders. The fine-tuned distillations meet or exceed the performance of previously published AI/ML solutions for each task. Further, the pipeline of distillation and model compression from TimesFM is generic and can be easily adapted to a variety of 1D waveform tasks across domains.