Title: A Hybrid Neural Network for High-Throughput Attosecond Resolution Single-shot X-ray Pulse Characterization Show Chinese title

Title: 一种用于高通量阿秒分辨率单次X射线脉冲表征的混合神经网络

Abstract: As scientific facilities transition toward high-throughput, data-intensive experiments, there is a growing need for real-time computing at the edge to support autonomous decision-making and minimize data transmission bottlenecks. Department of Energy (DOE) initiatives emphasize the development of heterogeneous computing architectures that integrate machine learning (ML) models with specialized hardware such as FPGAs, GPUs, and embedded systems to meet the demands of next-generation experiments. These advances are critical for facilities such as X-ray free-electron lasers (XFELs), where high-repetition-rate experiments produce terabyte-scale datasets, requiring real-time analysis and control. To address this challenge, we introduce DCIFR, a deep learning framework optimized for edge processing and high-speed X-ray diagnostics at SLAC's Linac Coherent Light Source (LCLS). DCIFR leverages a hybrid neural architecture, combining convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM) networks, and residual neural networks (ResNets) to denoise sinograms, classify X-ray sub-spikes, and extract phase separation with attosecond precision. The model achieves a phase extraction accuracy of 0.04652 radians (29.6 attoseconds at 1.2 um streaking wavelength) with inference latencies of 168.3 us and throughput exceeding 10kHz. Designed for deployment on heterogeneous architectures, DCIFR reduces computational overhead and supports future AI-driven feedback loops for ultrafast experiments. This work demonstrates the potential of edge AI and hardware acceleration to drive scalable, autonomous data analysis in next-generation DOE scientific facilities. Show Chinese abstract

Abstract: 随着科学设施向高通量、数据密集型实验转型，需要在边缘进行实时计算以支持自主决策并减少数据传输瓶颈。能源部（DOE）的计划强调开发异构计算架构，将机器学习（ML）模型与专用硬件如FPGAs、 GPUs和嵌入式系统相结合，以满足下一代实验的需求。这些进展对于X射线自由电子激光器（XFELs）等设施至关重要，其中高重复率实验会产生TB级数据集，需要实时分析和控制。为了解决这一挑战，我们介绍了DCIFR，这是一个专为SLAC直线加速器相干光源（LCLS）上的边缘处理和高速X射线诊断优化的深度学习框架。DCIFR利用混合神经架构，结合卷积神经网络（CNN）、双向长短期记忆（BiLSTM）网络和残差神经网络（ResNets），以实现对正弦图去噪、X射线子脉冲分类和飞秒精度的相位分离。该模型在推理延迟为168.3微秒的情况下实现了0.04652弧度的相位提取精度（在1.2微米条纹波长下为29.6飞秒），吞吐量超过10kHz。DCIFR设计用于部署在异构架构上，减少了计算开销，并支持未来超快实验的AI驱动反馈循环。这项工作展示了边缘AI和硬件加速在下一代DOE科学设施中推动可扩展、自主数据分析的潜力。