https://arxiv.org/abs/2112.13896

close this message
arXiv smileybones icon

Global Survey

In just 3 minutes, help us better understand how you perceive arXiv.
Take the survey

TAKE SURVEY
Skip to main content
Cornell University
We gratefully acknowledge support from
the Simons Foundation and member institutions.
 
arXiv.org > cs > arXiv:2112.13896
[                    ]

Help | Advanced Search

[All fields        ]
Search
arXiv
Cornell University Logo
[                    ] GO
quick links

  * Login
  * Help Pages
  * About

Computer Science > Machine Learning

arXiv:2112.13896 (cs)
[Submitted on 27 Dec 2021]

Title:Two Sparsities Are Better Than One: Unlocking the Performance
Benefits of Sparse-Sparse Networks

Authors:Kevin Lee Hunter, Lawrence Spracklen, Subutai Ahmad
Download PDF

    Abstract: In principle, sparse neural networks should be
    significantly more efficient than traditional dense networks.
    Neurons in the brain exhibit two types of sparsity; they are
    sparsely interconnected and sparsely active. These two types of
    sparsity, called weight sparsity and activation sparsity, when
    combined, offer the potential to reduce the computational cost of
    neural networks by two orders of magnitude. Despite this
    potential, today's neural networks deliver only modest
    performance benefits using just weight sparsity, because
    traditional computing hardware cannot efficiently process sparse
    networks. In this article we introduce Complementary Sparsity, a
    novel technique that significantly improves the performance of
    dual sparse networks on existing hardware. We demonstrate that we
    can achieve high performance running weight-sparse networks, and
    we can multiply those speedups by incorporating activation
    sparsity. Using Complementary Sparsity, we show up to 100X
    improvement in throughput and energy efficiency performing
    inference on FPGAs. We analyze scalability and resource tradeoffs
    for a variety of kernels typical of commercial convolutional
    networks such as ResNet-50 and MobileNetV2. Our results with
    Complementary Sparsity suggest that weight plus activation
    sparsity can be a potent combination for efficiently scaling
    future AI models.

Comments: 32 pages and 20 figures
          Machine Learning (cs.LG); Artificial Intelligence (cs.AI);
Subjects: Hardware Architecture (cs.AR); Neural and Evolutionary
          Computing (cs.NE)
Cite as:  arXiv:2112.13896 [cs.LG]
          (or arXiv:2112.13896v1 [cs.LG] for this version)

Submission history

From: Subutai Ahmad [view email]
[v1] Mon, 27 Dec 2021 20:41:01 UTC (8,764 KB)
Full-text links:

Download:

  * PDF
  * Other formats

[by-4]
Current browse context:
cs.LG
< prev   |   next >
new | recent | 2112
Change to browse by:
cs
cs.AI
cs.AR
cs.NE

References & Citations

  * NASA ADS
  * Google Scholar
  * Semantic Scholar

DBLP - CS Bibliography

listing | bibtex
Subutai Ahmad
a export bibtex citation Loading...

Bibtex formatted citation

x
[loading...          ]
Data provided by:

Bookmark

BibSonomy logo Mendeley logo Reddit logo ScienceWISE logo
(*) Bibliographic Tools

Bibliographic and Citation Tools

[ ] Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
[ ] Litmaps Toggle
Litmaps (What is Litmaps?)
[ ] scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
( ) Code & Data

Code and Data Associated with this Article

[ ] arXiv Links to Code Toggle
arXiv Links to Code & Data (What is Links to Code & Data?)
( ) Related Papers

Recommenders and Search Tools

[ ] Connected Papers Toggle
Connected Papers (What is Connected Papers?)
[ ] Core recommender toggle
CORE Recommender (What is CORE?)
( ) About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and
share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have
embraced and accepted our values of openness, community, excellence,
and user data privacy. arXiv is committed to these values and only
works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community?
Learn more about arXivLabs and how to get involved.

Which authors of this paper are endorsers? | Disable MathJax (What is
MathJax?)

  * About
  * Help

  * Click here to contact arXiv Contact
  * Click here to subscribe Subscribe

  * Copyright
  * Privacy Policy

  * Web Accessibility Assistance
  * arXiv Operational Status
    Get status notifications via email or slack