DARL: Distributed Reconfigurable Accelerator for Hyperdimensional Reinforcement Learning

Reinforcement Learning (RL) is a powerful technology to solve decision-making problems such as robotics control. Modern RL algorithms, i.e., Deep Q-Learning, are based on costly and resource hungry deep neural networks. This motivates us to deploy alternative models for powering RL agents on edge de...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2022 IEEE/ACM International Conference On Computer Aided Design (ICCAD) S. 1 - 9
Hauptverfasser: Chen, Hanning, Issa, Mariam, Ni, Yang, Imani, Mohsen
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: ACM 29.10.2022
Schlagworte:
ISSN:1558-2434
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Reinforcement Learning (RL) is a powerful technology to solve decision-making problems such as robotics control. Modern RL algorithms, i.e., Deep Q-Learning, are based on costly and resource hungry deep neural networks. This motivates us to deploy alternative models for powering RL agents on edge devices. Recently, brain-inspired Hyper-Dimensional Computing (HDC) has been introduced as a promising solution for lightweight and efficient machine learning, particularly for classification.In this work, we develop a novel platform capable of real-time hyper-dimensional reinforcement learning. Our heterogeneous CPU-FPGA platform, called DARL, maximizes FPGA's computing capabilities by applying hardware optimizations to hyperdimensional computing's critical operations, including hardware-friendly encoder IP, the hypervector chunk fragmentation, and the delayed model update. Aside from hardware innovation, we also extend the platform to basic single-agent RL to support multi-agents distributed learning. We evaluate the effectiveness of our approach on OpenAI Gym tasks. Our results show that the FPGA platform provides on average 20× speedup compared to current state-of-the-art hyperdimensional RL methods running on Intel Xeon 6226 CPU. In addition, DARL provides around 4.8× faster and 4.2× higher energy efficiency compared to the state-of-the-art RL accelerator while ensuring a better or comparable quality of learning.
AbstractList Reinforcement Learning (RL) is a powerful technology to solve decision-making problems such as robotics control. Modern RL algorithms, i.e., Deep Q-Learning, are based on costly and resource hungry deep neural networks. This motivates us to deploy alternative models for powering RL agents on edge devices. Recently, brain-inspired Hyper-Dimensional Computing (HDC) has been introduced as a promising solution for lightweight and efficient machine learning, particularly for classification.In this work, we develop a novel platform capable of real-time hyper-dimensional reinforcement learning. Our heterogeneous CPU-FPGA platform, called DARL, maximizes FPGA's computing capabilities by applying hardware optimizations to hyperdimensional computing's critical operations, including hardware-friendly encoder IP, the hypervector chunk fragmentation, and the delayed model update. Aside from hardware innovation, we also extend the platform to basic single-agent RL to support multi-agents distributed learning. We evaluate the effectiveness of our approach on OpenAI Gym tasks. Our results show that the FPGA platform provides on average 20× speedup compared to current state-of-the-art hyperdimensional RL methods running on Intel Xeon 6226 CPU. In addition, DARL provides around 4.8× faster and 4.2× higher energy efficiency compared to the state-of-the-art RL accelerator while ensuring a better or comparable quality of learning.
Author Ni, Yang
Chen, Hanning
Imani, Mohsen
Issa, Mariam
Author_xml – sequence: 1
  givenname: Hanning
  surname: Chen
  fullname: Chen, Hanning
  email: hanningc@uci.edu
  organization: University of California,Department of Computer Science,Irvine,CA,USA
– sequence: 2
  givenname: Mariam
  surname: Issa
  fullname: Issa, Mariam
  email: mariamai@uci.edu
  organization: University of California,Department of Computer Science,Irvine,CA,USA
– sequence: 3
  givenname: Yang
  surname: Ni
  fullname: Ni, Yang
  email: yni3@uci.edu
  organization: University of California,Department of Computer Science,Irvine,CA,USA
– sequence: 4
  givenname: Mohsen
  surname: Imani
  fullname: Imani, Mohsen
  email: m.imani@uci.edu
  organization: University of California,Department of Computer Science,Irvine,CA,USA
BookMark eNotj01Lw0AYhFdRsNacvXjIH0h993vjrbTVCgGhKHgru5s3ZSXdlE166L93QQ_DwDzDwNyTmzhEJOSRwoJSIZ-5BMMlW3ApasH1FSlqbTIAXjOqxTWZUSlNxQQXd6QYxx8AYEZTrWFGvtfLXfNSrsM4peDOE7blDv0Qu3A4J-t6LJfeY4_JTkMqu6zt5YSpDUeMYxii7XM_xAw85mgqG7Qphnh4ILed7Ucs_n1Ovl43n6tt1Xy8va-WTWWZMFMlgKHvGPeoZAud14qJmkmknDmtvUJv2xadMdQBze-UswIUB_A1ExQZn5Onv92AiPtTCkebLnsKoExNDf8FxlxUQw
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1145/3508352.3549437
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISBN 9781450392174
1450392172
EISSN 1558-2434
EndPage 9
ExternalDocumentID 10068918
Genre orig-research
GrantInformation_xml – fundername: Air Force Office of Scientific Research
  funderid: 10.13039/100000181
– fundername: Semiconductor Research Corporation
  funderid: 10.13039/100000028
– fundername: Office of Naval Research
  funderid: 10.13039/100000006
– fundername: National Science Foundation
  funderid: 10.13039/100000001
GroupedDBID 6IE
6IF
6IH
6IL
6IN
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
FEDTE
IEGSK
IJVOP
M43
OCL
RIE
RIL
RIO
ID FETCH-LOGICAL-a248t-402ecf23ce65d0fc7624925e132b77c6ecaddeb881b014376ba406300c9241e23
IEDL.DBID RIE
ISICitedReferencesCount 17
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000981574300083&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 02:46:18 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a248t-402ecf23ce65d0fc7624925e132b77c6ecaddeb881b014376ba406300c9241e23
PageCount 9
ParticipantIDs ieee_primary_10068918
PublicationCentury 2000
PublicationDate 2022-Oct.-29
PublicationDateYYYYMMDD 2022-10-29
PublicationDate_xml – month: 10
  year: 2022
  text: 2022-Oct.-29
  day: 29
PublicationDecade 2020
PublicationTitle 2022 IEEE/ACM International Conference On Computer Aided Design (ICCAD)
PublicationTitleAbbrev ICCAD
PublicationYear 2022
Publisher ACM
Publisher_xml – name: ACM
SSID ssj0002871770
ssj0020286
Score 2.3460808
Snippet Reinforcement Learning (RL) is a powerful technology to solve decision-making problems such as robotics control. Modern RL algorithms, i.e., Deep Q-Learning,...
SourceID ieee
SourceType Publisher
StartPage 1
SubjectTerms Computational modeling
Computer aided instruction
Distance learning
Hardware
IP networks
Real-time systems
Technological innovation
Title DARL: Distributed Reconfigurable Accelerator for Hyperdimensional Reinforcement Learning
URI https://ieeexplore.ieee.org/document/10068918
WOSCitedRecordID wos000981574300083&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV09T8MwELVoxQALX0V8ywNrShonvpitolQdqqpCIHWrYvtSdSBFpeX3c-eGAgMDW2RlsPx175393glxC9Z6mxJyU6pwUZpaiGyeEZArwVttKGDa4DM7hNEon0zMuBarBy0MIobHZ9jmz3CX7xduzaky2uGxzk0nb4gGAGzEWtuECkN_4MVXsy1q0LWXTyfN7lQWwEZbESFK1e9iKiGW9A_-2YtD0fpW5cnxNt4ciR2sjsX-D0PBEzHpdZ-G97LHbrhcyAq9ZH5ZlfPZeskiKdl1jgJNuFuXhFflgHgorZFXfsfOoJz-D16qLqQNZW2_OmuJl_7j88MgqmsnREWS5iumhejKRDnUmY9LR2ce2xAikU8L4DQ6PthsTqiVHf5A2yIN9luOCFkHE3UqmtWiwjMhTay98UaDJ7jlYyhUZhBj0MaXHnJzLlo8SNO3jT3G9Gt8Lv5ovxR7CWsIKAAk5ko0V8s1Xotd97Gavy9vwqR-ApRLohs
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV27TsMwFLWgIAELryLeeGBNSWLHjtkqSlVEqCpUpG5V_EjVgRSFlu_nXjcUGBjYIiuD5dc959rnXEKupdZWc0BujOUm4FzLQKcJALlCWi0UBEztfWYz2e-no5Ea1GJ1r4VxzvnHZ66Fn_4u387MAlNlsMNDkaooXScbCedxtJRrrVIqCP4lLr-ab0GDqN18Ip7csMTDjRYDSsTZ73IqPpp0d__Zjz3S_Nbl0cEq4uyTNVcekJ0floKHZNRpP2e3tIN-uFjKylmKDLMsppNFhTIp2jYGQo2_XaeAWGkPmCiskld8yY6wHP73bqrGJw5pbcA6aZKX7v3wrhfU1ROCPObpHImhM0XMjBOJDQsDpx4aETqgn1pKI5zBo02ngFvR408KnXNvwGWAkkUuZkekUc5Kd0yoCoVVVglpAXDZUOYsUc6FUihbWJmqE9LEQRq_LQ0yxl_jc_pH-xXZ6g2fsnH20H88I9sxKgogHMTqnDTm1cJdkE3zMZ--V5d-gj8BMu-lYg
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=2022+IEEE%2FACM+International+Conference+On+Computer+Aided+Design+%28ICCAD%29&rft.atitle=DARL%3A+Distributed+Reconfigurable+Accelerator+for+Hyperdimensional+Reinforcement+Learning&rft.au=Chen%2C+Hanning&rft.au=Issa%2C+Mariam&rft.au=Ni%2C+Yang&rft.au=Imani%2C+Mohsen&rft.date=2022-10-29&rft.pub=ACM&rft.eissn=1558-2434&rft.spage=1&rft.epage=9&rft_id=info:doi/10.1145%2F3508352.3549437&rft.externalDocID=10068918