close this message
arXiv smileybones icon

Giving Week!

Show your support for Open Science by donating to arXiv during Giving Week, April 25th-29th.

DONATE
Skip to main content
Cornell University
We gratefully acknowledge support from
the Simons Foundation and member institutions.
arxiv logo > cs > arXiv:2202.01994

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Machine Learning

arXiv:2202.01994 (cs)
[Submitted on 4 Feb 2022]

Title:Data Scaling Laws in NMT: The Effect of Noise and Architecture

Authors:Yamini Bansal, Behrooz Ghorbani, Ankush Garg, Biao Zhang, Maxim Krikun, Colin Cherry, Behnam Neyshabur, Orhan Firat
Download PDF
Abstract: In this work, we study the effect of varying the architecture and training data quality on the data scaling properties of Neural Machine Translation (NMT). First, we establish that the test loss of encoder-decoder transformer models scales as a power law in the number of training samples, with a dependence on the model size. Then, we systematically vary aspects of the training setup to understand how they impact the data scaling laws. In particular, we change the following (1) Architecture and task setup: We compare to a transformer-LSTM hybrid, and a decoder-only transformer with a language modeling loss (2) Noise level in the training distribution: We experiment with filtering, and adding iid synthetic noise. In all the above cases, we find that the data scaling exponents are minimally impacted, suggesting that marginally worse architectures or training data can be compensated for by adding more data. Lastly, we find that using back-translated data instead of parallel data, can significantly degrade the scaling exponent.
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Cite as: arXiv:2202.01994 [cs.LG]
  (or arXiv:2202.01994v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2202.01994
arXiv-issued DOI via DataCite

Submission history

From: Yamini Bansal [view email]
[v1] Fri, 4 Feb 2022 06:53:49 UTC (2,641 KB)
Full-text links:

Download:

  • PDF
  • Other formats
(license)
Current browse context:
cs.LG
< prev   |   next >
new | recent | 2202
Change to browse by:
cs
cs.CL

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar

DBLP - CS Bibliography

listing | bibtex
Yamini Bansal
Behrooz Ghorbani
Ankush Garg
Biao Zhang
Maxim Krikun
…
a export bibtex citation Loading...

Bibtex formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Mendeley logo Reddit logo ScienceWISE logo

Bibliographic and Citation Tools

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

Code and Data Associated with this Article

arXiv Links to Code & Data (What is Links to Code & Data?)

Demos

Replicate (What is Replicate?)

Recommenders and Search Tools

Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)

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
  • 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