A Hardware Acceleration Platform for AI-Based Inference at the Edge

Kimon Karras, Evangelos Pallis, George Mastorakis, Yannis Nikoloudakis, Jordi Mongay Batalla, Constandinos X. Mavromoustakis, Evangelos Markakis

    Research output: Contribution to journalArticle

    Abstract

    Machine learning (ML) algorithms are already transforming the way data are collected and processed in the data center, where some form of AI has permeated most areas of computing. The integration of AI algorithms at the edge is the next logical step which is already under investigation. However, harnessing such algorithms at the edge will require more computing power than what current platforms offer. In this paper, we present an FPGA system-on-chip-based architecture that supports the acceleration of ML algorithms in an edge environment. The system supports dynamic deployment of ML functions driven either locally or remotely, thus achieving a remarkable degree of flexibility. We demonstrate the efficacy of this architecture by executing a version of the well-known YOLO classifier which demonstrates competitive performance while requiring a reasonable amount of resources on the device.

    Original languageEnglish
    JournalCircuits, Systems, and Signal Processing
    DOIs
    Publication statusPublished - 1 Jan 2019

    Fingerprint

    Hardware Acceleration
    Learning systems
    Machine Learning
    Hardware
    Learning algorithms
    Learning Algorithm
    Computing
    Data Center
    Demonstrate
    Field Programmable Gate Array
    Field programmable gate arrays (FPGA)
    Efficacy
    Classifiers
    Flexibility
    Classifier
    Resources
    Architecture

    Keywords

    • Acceleration
    • Acceleration of machine learning
    • AI
    • Computing
    • EDGE
    • Fog
    • ML
    • PCP
    • YOLO

    Cite this

    Karras, Kimon ; Pallis, Evangelos ; Mastorakis, George ; Nikoloudakis, Yannis ; Batalla, Jordi Mongay ; Mavromoustakis, Constandinos X. ; Markakis, Evangelos. / A Hardware Acceleration Platform for AI-Based Inference at the Edge. In: Circuits, Systems, and Signal Processing. 2019.
    @article{a9220807937d4f27a4e375106673ef0e,
    title = "A Hardware Acceleration Platform for AI-Based Inference at the Edge",
    abstract = "Machine learning (ML) algorithms are already transforming the way data are collected and processed in the data center, where some form of AI has permeated most areas of computing. The integration of AI algorithms at the edge is the next logical step which is already under investigation. However, harnessing such algorithms at the edge will require more computing power than what current platforms offer. In this paper, we present an FPGA system-on-chip-based architecture that supports the acceleration of ML algorithms in an edge environment. The system supports dynamic deployment of ML functions driven either locally or remotely, thus achieving a remarkable degree of flexibility. We demonstrate the efficacy of this architecture by executing a version of the well-known YOLO classifier which demonstrates competitive performance while requiring a reasonable amount of resources on the device.",
    keywords = "Acceleration, Acceleration of machine learning, AI, Computing, EDGE, Fog, ML, PCP, YOLO",
    author = "Kimon Karras and Evangelos Pallis and George Mastorakis and Yannis Nikoloudakis and Batalla, {Jordi Mongay} and Mavromoustakis, {Constandinos X.} and Evangelos Markakis",
    year = "2019",
    month = "1",
    day = "1",
    doi = "10.1007/s00034-019-01226-7",
    language = "English",
    journal = "Circuits, Systems, and Signal Processing",
    issn = "0278-081X",

    }

    A Hardware Acceleration Platform for AI-Based Inference at the Edge. / Karras, Kimon; Pallis, Evangelos; Mastorakis, George; Nikoloudakis, Yannis; Batalla, Jordi Mongay; Mavromoustakis, Constandinos X.; Markakis, Evangelos.

    In: Circuits, Systems, and Signal Processing, 01.01.2019.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - A Hardware Acceleration Platform for AI-Based Inference at the Edge

    AU - Karras, Kimon

    AU - Pallis, Evangelos

    AU - Mastorakis, George

    AU - Nikoloudakis, Yannis

    AU - Batalla, Jordi Mongay

    AU - Mavromoustakis, Constandinos X.

    AU - Markakis, Evangelos

    PY - 2019/1/1

    Y1 - 2019/1/1

    N2 - Machine learning (ML) algorithms are already transforming the way data are collected and processed in the data center, where some form of AI has permeated most areas of computing. The integration of AI algorithms at the edge is the next logical step which is already under investigation. However, harnessing such algorithms at the edge will require more computing power than what current platforms offer. In this paper, we present an FPGA system-on-chip-based architecture that supports the acceleration of ML algorithms in an edge environment. The system supports dynamic deployment of ML functions driven either locally or remotely, thus achieving a remarkable degree of flexibility. We demonstrate the efficacy of this architecture by executing a version of the well-known YOLO classifier which demonstrates competitive performance while requiring a reasonable amount of resources on the device.

    AB - Machine learning (ML) algorithms are already transforming the way data are collected and processed in the data center, where some form of AI has permeated most areas of computing. The integration of AI algorithms at the edge is the next logical step which is already under investigation. However, harnessing such algorithms at the edge will require more computing power than what current platforms offer. In this paper, we present an FPGA system-on-chip-based architecture that supports the acceleration of ML algorithms in an edge environment. The system supports dynamic deployment of ML functions driven either locally or remotely, thus achieving a remarkable degree of flexibility. We demonstrate the efficacy of this architecture by executing a version of the well-known YOLO classifier which demonstrates competitive performance while requiring a reasonable amount of resources on the device.

    KW - Acceleration

    KW - Acceleration of machine learning

    KW - AI

    KW - Computing

    KW - EDGE

    KW - Fog

    KW - ML

    KW - PCP

    KW - YOLO

    UR - http://www.scopus.com/inward/record.url?scp=85070324520&partnerID=8YFLogxK

    U2 - 10.1007/s00034-019-01226-7

    DO - 10.1007/s00034-019-01226-7

    M3 - Article

    JO - Circuits, Systems, and Signal Processing

    JF - Circuits, Systems, and Signal Processing

    SN - 0278-081X

    ER -