Skip to main content
Cornell University
We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate
arxiv logo > cs > arXiv:1711.02257

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Computer Vision and Pattern Recognition

arXiv:1711.02257 (cs)
[Submitted on 7 Nov 2017 (v1), last revised 12 Jun 2018 (this version, v4)]

Title:GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks

Authors:Zhao Chen, Vijay Badrinarayanan, Chen-Yu Lee, Andrew Rabinovich
View a PDF of the paper titled GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks, by Zhao Chen and 2 other authors
View PDF
Abstract:Deep multitask networks, in which one neural network produces multiple predictive outputs, can offer better speed and performance than their single-task counterparts but are challenging to train properly. We present a gradient normalization (GradNorm) algorithm that automatically balances training in deep multitask models by dynamically tuning gradient magnitudes. We show that for various network architectures, for both regression and classification tasks, and on both synthetic and real datasets, GradNorm improves accuracy and reduces overfitting across multiple tasks when compared to single-task networks, static baselines, and other adaptive multitask loss balancing techniques. GradNorm also matches or surpasses the performance of exhaustive grid search methods, despite only involving a single asymmetry hyperparameter $\alpha$. Thus, what was once a tedious search process that incurred exponentially more compute for each task added can now be accomplished within a few training runs, irrespective of the number of tasks. Ultimately, we will demonstrate that gradient manipulation affords us great control over the training dynamics of multitask networks and may be one of the keys to unlocking the potential of multitask learning.
Comments: ICML 2018
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:1711.02257 [cs.CV]
  (or arXiv:1711.02257v4 [cs.CV] for this version)
  https://doi.org/10.48550/arXiv.1711.02257
arXiv-issued DOI via DataCite
Journal reference: Proceedings of the 35th International Conference on Machine Learning (2018), 793-802

Submission history

From: Zhao Chen [view email]
[v1] Tue, 7 Nov 2017 02:08:12 UTC (1,275 KB)
[v2] Tue, 19 Dec 2017 01:00:22 UTC (1,278 KB)
[v3] Sun, 8 Apr 2018 21:25:49 UTC (1,874 KB)
[v4] Tue, 12 Jun 2018 06:45:49 UTC (1,650 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks, by Zhao Chen and 2 other authors
  • View PDF
  • TeX Source
  • Other Formats
view license
Current browse context:
cs.CV
< prev   |   next >
new | recent | 2017-11
Change to browse by:
cs

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar

1 blog link

(what is this?)

DBLP - CS Bibliography

listing | bibtex
Zhao Chen
Vijay Badrinarayanan
Chen-Yu Lee
Andrew Rabinovich
a export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)

Demos

Replicate (What is Replicate?)
Hugging Face Spaces (What is Spaces?)
TXYZ.AI (What is TXYZ.AI?)

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
  • Author
  • Venue
  • Institution
  • Topic

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.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)
  • About
  • Help
  • contact arXivClick here to contact arXiv Contact
  • subscribe to arXiv mailingsClick here to subscribe Subscribe
  • Copyright
  • Privacy Policy
  • Web Accessibility Assistance
  • arXiv Operational Status
    Get status notifications via email or slack