Related papers: The Fast Linear Accelerator Modeling Engine for FR…

The rapid adaptation of data driven AI models, such as deep learning inference, training, Vision Transformers (ViTs), and other HPC applications, drives a strong need for runtime precision configurable different non linear activation…

Large language models (LLMs) have emerged as a powerful foundation for intelligent reasoning and decision-making, demonstrating substantial impact across a wide range of domains and applications. However, their massive parameter scales and…

Text-based motion generation models are drawing a surge of interest for their potential for automating the motion-making process in the game, animation, or robot industries. In this paper, we propose a diffusion-based motion synthesis and…

Fermilab is planning the construction of a prototype front end of the Project X linac. The Project X Injector Experiment (PXIE) is expected to accelerate 1 mA cw H- beam up to 30 MeV. Some of the major goals of the project are to test a cw…

Neural network accelerators with low latency and low energy consumption are desirable for edge computing. To create such accelerators, we propose a design flow for accelerating the extremely low bit-width neural network (ELB-NN) in embedded…

In this work, we introduce FLAME, a family of extremely lightweight and capable Time Series Foundation Models, which support both deterministic and probabilistic forecasting via generative probabilistic modeling, thus ensuring both…

Laser plasma accelerators have the potential to reduce the size of future linacs for high energy physics by more than an order of magnitude, due to their high gradient. Research is in progress at current facilities, including the BELLA…

Flow and diffusion models achieve high-fidelity, high-resolution image synthesis, but often require many function evaluations (NFEs) at sampling time. Existing acceleration methods either require additional training through distillation or…

Wireless federated learning (WFL) suffers from heterogeneity prevailing in the data distributions, computing powers, and channel conditions of participating devices. This paper presents a new Federated Learning with Adjusted leaRning ratE…

The current program at Fermilab involves the construction of a new superconducting linear accelerator (LINAC) to replace the existing warm version. The new LINAC, together with other planned improvements, is in support of proton beam…

With the rapid development of large language models (LLM), robots are starting to enjoy the benefits of new interaction methods that large language models bring. Because edge computing fulfills the needs for rapid response, privacy, and…

The CLIC study of a high-energy (0.5 - 5 TeV), high-luminosity (1034 - 1035 cm-2 sec-1) e+e- linear collider is presented. Beam acceleration using high frequency (30 GHz) normal-conducting structures operating at high accelerating fields…

This paper presents a machine learning-accelerated optimization framework for RF power amplifier design that reduces simulation requirements by 65% while maintaining $\pm0.4$ dBm accuracy for the majority of the modes. The proposed method…

As large language models (LLMs) advance, it becomes more challenging to reliably evaluate their output due to the high costs of human evaluation. To make progress towards better LLM autoraters, we introduce FLAMe, a family of Foundational…

Large language model (LLM) pruning with fixed N:M structured sparsity significantly limits the expressivity of the sparse model, yielding sub-optimal performance. In contrast, supporting multiple N:M patterns to provide sparse…

Modeling of laser-plasma wakefield accelerators in an optimal frame of reference \cite{VayPRL07} is shown to produce orders of magnitude speed-up of calculations from first principles. Obtaining these speedups requires mitigation of a…

The performance of deep learning models is critically dependent on sophisticated optimization strategies. While existing optimizers have shown promising results, many rely on first-order Exponential Moving Average (EMA) techniques, which…

The Fermilab Linac delivers 400 MeV H- beam to the rest of the accelerator chain. Providing stable intensity, energy, and emittance is key since it directly affects downstream machines. To counter fluctuations of Linac output due to various…

In the emerging paradigm of Federated Learning (FL), large amount of clients such as mobile devices are used to train possibly high-dimensional models on their respective data. Combining (dimension-wise) adaptive gradient methods (e.g.…

Numerical simulation of the National Ignition Facility (NIF) laser performance and automated control of the laser setup process are crucial to the project's success. These functions will be performed by two closely coupled computer code:…