hls4ml: An Open-Source Codesign Workflow to Empower Scientific Low-Power Machine Learning Devices

Accessible machine learning algorithms, software, and diagnostic tools for energy-efficient devices and systems are extremely valuable across a broad range of application domains. In scientific domains, real-time near-sensor processing can drastically improve experimental design and accelerate scien...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2021-03
Main Authors: Fahim, Farah, Hawks, Benjamin, Herwig, Christian, Hirschauer, James, Sergo Jindariani, Tran, Nhan, Carloni, Luca P, Giuseppe Di Guglielmo, Harris, Philip, Krupa, Jeffrey, Rankin, Dylan, Manuel Blanco Valentin, Hester, Josiah, Luo, Yingyi, Mamish, John, Orgrenci-Memik, Seda, Aarrestad, Thea, Hamza Javed, Loncar, Vladimir, Pierini, Maurizio, Pol, Adrian Alan, Summers, Sioni, Duarte, Javier, Hauck, Scott, Hsu, Shih-Chieh, Ngadiuba, Jennifer, Liu, Mia, Hoang, Duc, Kreinar, Edward, Wu, Zhenbin
Format: Article
Language:English
Subjects:
Accessibility
Algorithms
Application specific integrated circuits
Co-design
Design of experiments
Domains
Machine learning
Open source software
Power
Power management
Scientists
Workflow
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Accessible machine learning algorithms, software, and diagnostic tools for energy-efficient devices and systems are extremely valuable across a broad range of application domains. In scientific domains, real-time near-sensor processing can drastically improve experimental design and accelerate scientific discoveries. To support domain scientists, we have developed hls4ml, an open-source software-hardware codesign workflow to interpret and translate machine learning algorithms for implementation with both FPGA and ASIC technologies. We expand on previous hls4ml work by extending capabilities and techniques towards low-power implementations and increased usability: new Python APIs, quantization-aware pruning, end-to-end FPGA workflows, long pipeline kernels for low power, and new device backends include an ASIC workflow. Taken together, these and continued efforts in hls4ml will arm a new generation of domain scientists with accessible, efficient, and powerful tools for machine-learning-accelerated discovery.
ISSN:2331-8422