Search NASA⌕ Search
Home/Archive/DOE OSTI

DOE OSTI · 3022473

hls4ml: A Flexible, Open-Source Platform for Deep Learning Acceleration on Reconfigurable Hardware

Schulte, Jan-Frederik·Ramhorst, Benjamin·Sun, Chang·Mitrevski, Jovan·Ghielmetti, Nicolò·Lupi, Enrico·Danopoulos, Dimitrios·Loncar, Vladimir·Duarte, Javier·Burnette, David·Laatu, Lauri·Tzelepis, Stylianos·Axiotis, Konstantinos·Berthet, Quentin·Wang, Haoyan·White, Paul·Demirsoy, Suleyman·Colombo, Marco·Aarrestad, Thea·Summers, Sioni·Pierini, Maurizio·Di Guglielmo, Giuseppe·Ngadiuba, Jennifer·Campos, Javier·Hawks, Ben·Gandrakota, Abhijith·Fahim, Farah·Tran, Nhan·Constantinides, George·Que, Zhiqiang·Luk, Wayne·Tapper, Alexander·Hoang, Duc·Paladino, Noah·Harris, Philip·Lai, Bo-Cheng·Valentin, Manuel·Forelli, Ryan·Ogrenci, Seda·Gerlach, Lino·Flynn, Rian·Liu, Mia·Diaz, Daniel·Khoda, Elham·Quinnan, Melissa·Solares, Russell·Parajuli, Santosh·Neubauer, Mark·Herwig, Christian·Tsoi, Ho Fung·Rankin, Dylan·Hsu, Shih-Chieh·Hauck, Scott
View original source ↗Read source document ↗Export metadata ↓

Abstract

We present hls4ml, a free and open-source platform that translates machine learning (ML) models from modern deep learning frameworks into high-level synthesis (HLS) code that can be integrated into full designs for field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs). With its flexible and modular design, hls4ml supports a large number of deep learning frameworks and can target HLS compilers from several vendors, including Vitis HLS, Intel oneAPI and Catapult HLS. Together with a wider eco-system for software-hardware co-design, hls4ml has enabled the acceleration of ML inference in a wide range of commercial and scientific applications where low latency, resource usage, and power consumption are critical. In this paper, we describe the structure and functionality of the hls4ml platform. The overarching design considerations for the generated HLS code are discussed, together with selected performance results.

Explore related subjects

FOS: Computer and information sciencesFOS: Physical sciencesHardware Architecture (cs.AR)High Energy Physics - Experiment (hep-ex)Machine Learning (cs.LG)

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Schulte, Jan-Frederik, Ramhorst, Benjamin, Sun, Chang, Mitrevski, Jovan, Ghielmetti, Nicolò, Lupi, Enrico, Danopoulos, Dimitrios, Loncar, Vladimir, Duarte, Javier, Burnette, David, Laatu, Lauri, Tzelepis, Stylianos, Axiotis, Konstantinos, Berthet, Quentin, Wang, Haoyan, White, Paul, Demirsoy, Suleyman, Colombo, Marco, Aarrestad, Thea, Summers, Sioni, Pierini, Maurizio, Di Guglielmo, Giuseppe, Ngadiuba, Jennifer, Campos, Javier, Hawks, Ben, Gandrakota, Abhijith, Fahim, Farah, Tran, Nhan, Constantinides, George, Que, Zhiqiang, Luk, Wayne, Tapper, Alexander, Hoang, Duc, Paladino, Noah, Harris, Philip, Lai, Bo-Cheng, Valentin, Manuel, Forelli, Ryan, Ogrenci, Seda, Gerlach, Lino, Flynn, Rian, Liu, Mia, Diaz, Daniel, Khoda, Elham, Quinnan, Melissa, Solares, Russell, Parajuli, Santosh, Neubauer, Mark, Herwig, Christian, Tsoi, Ho Fung, Rankin, Dylan, Hsu, Shih-Chieh, Hauck, Scott. 2025-01-01. hls4ml: A Flexible, Open-Source Platform for Deep Learning Acceleration on Reconfigurable Hardware. https://doi.org/10.48550/arxiv.2512.01463

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

DOE OSTI2026

Domain Knowledge Guided Bayesian Optimization For Autonomous Alignment Of Complex Scientific Instruments

Bayesian Optimization (BO) is a powerful tool for optimizing complex non-linear systems. However, its performance degrades in high-dimensional problems with tightly coupled parameters and highly asymmetric objective landscapes, where rewards are sparse. In such needle-in-a-haystack scenarios, even advanced methods like trust-region BO (TurBO) often lead to unsatisfactory results. We propose a domain knowledge guided Bayesian Optimization approach, which leverages physical insight to fundamentally simplify the search problem by transforming coordinates to decouple input features and align the active subspaces with the primary search axes. We demonstrate this approach's efficacy on a challenging 12-dimensional, 6-crystal Split-and-Delay optical system, where conventional approaches, including standard BO, TuRBO and multi-objective BO, consistently led to unsatisfactory results. When combined with an reverse annealing exploration strategy, this approach reliably converges to the global optimum. The coordinate transformation itself is the key to this success, significantly accelerating the search by aligning input co-ordinate axes with the problem's active subspaces. As increasingly complex scientific instruments, from large telescopes to new spectrometers at X-ray Free Electron Lasers are deployed, the demand for robust high-dimensional optimization grows. Our results demonstrate a generalizable paradigm: leveraging physical insight to transform high-dimensional, coupled optimization problems into simpler representations can enable rapid and robust automated tuning for consistent high performance while still retaining current optimization algorithms.

FOS: Computer and information sciences↗
DOE OSTI2025

Global Framework for Emulation of Nuclear Calculations

We introduce a hierarchical framework that combines ab initio many-body calculations with a Bayesian neural network, developing emulators capable of accurately predicting nuclear properties across isotopic chains simultaneously and being applicable to different regions of the nuclear chart. We benchmark our developments using the oxygen isotopic chain, achieving accurate results for ground-state energies and nuclear charge radii, while providing robust uncertainty quantification. Our framework enables global sensitivity analysis of nuclear binding energies and charge radii with respect to the low-energy constants that describe the nuclear force.

FOS: Computer and information sciences↗
DOE OSTI2025

Weak Form Scientific Machine Learning: Test Function Construction for System Identification

Weak form Scientific Machine Learning (WSciML) is a recently developed framework for data-driven modeling and scientific discovery. It leverages the weak form of equation error residuals to provide enhanced noise robustness in system identification via convolving model equations with test functions, reformulating the problem to avoid direct differentiation of data. The performance, however, relies on wisely choosing a set of compactly supported test functions. In this work, we mathematically motivate a novel data-driven method for constructing Single-scale-Local reference functions for creating the set of test functions. Our approach numerically approximates the integration error introduced by the quadrature and identifies the support size for which the error is minimal, without requiring access to the model parameter values. Through numerical experiments across various models, noise levels, and temporal resolutions, we demonstrate that the selected supports consistently align with regions of minimal parameter estimation error. We also compare the proposed method against the strategy for constructing Multi-scale-Global (and orthogonal) test functions introduced in our prior work, demonstrating the improved computational efficiency.

FOS: Computer and information sciences↗