Skip to main content
Cornell University
We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate
arxiv logo > stat > arXiv:1804.09060

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Statistics > Machine Learning

arXiv:1804.09060 (stat)
[Submitted on 24 Apr 2018 (v1), last revised 2 Oct 2018 (this version, v8)]

Title:An Information-Theoretic View for Deep Learning

Authors:Jingwei Zhang, Tongliang Liu, Dacheng Tao
View a PDF of the paper titled An Information-Theoretic View for Deep Learning, by Jingwei Zhang and 2 other authors
View PDF
Abstract:Deep learning has transformed computer vision, natural language processing, and speech recognition\cite{badrinarayanan2017segnet, dong2016image, ren2017faster, ji20133d}. However, two critical questions remain obscure: (1) why do deep neural networks generalize better than shallow networks; and (2) does it always hold that a deeper network leads to better performance? Specifically, letting $L$ be the number of convolutional and pooling layers in a deep neural network, and $n$ be the size of the training sample, we derive an upper bound on the expected generalization error for this network, i.e.,
\begin{eqnarray*}
\mathbb{E}[R(W)-R_S(W)] \leq \exp{\left(-\frac{L}{2}\log{\frac{1}{\eta}}\right)}\sqrt{\frac{2\sigma^2}{n}I(S,W) }
\end{eqnarray*} where $\sigma >0$ is a constant depending on the loss function, $0<\eta<1$ is a constant depending on the information loss for each convolutional or pooling layer, and $I(S, W)$ is the mutual information between the training sample $S$ and the output hypothesis $W$. This upper bound shows that as the number of convolutional and pooling layers $L$ increases in the network, the expected generalization error will decrease exponentially to zero. Layers with strict information loss, such as the convolutional layers, reduce the generalization error for the whole network; this answers the first question. However, algorithms with zero expected generalization error does not imply a small test error or $\mathbb{E}[R(W)]$. This is because $\mathbb{E}[R_S(W)]$ is large when the information for fitting the data is lost as the number of layers increases. This suggests that the claim `the deeper the better' is conditioned on a small training error or $\mathbb{E}[R_S(W)]$. Finally, we show that deep learning satisfies a weak notion of stability and the sample complexity of deep neural networks will decrease as $L$ increases.
Comments: Add details in the proof of Theorem 2
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:1804.09060 [stat.ML]
  (or arXiv:1804.09060v8 [stat.ML] for this version)
  https://doi.org/10.48550/arXiv.1804.09060
arXiv-issued DOI via DataCite

Submission history

From: Dacheng Tao [view email]
[v1] Tue, 24 Apr 2018 14:13:19 UTC (398 KB)
[v2] Thu, 3 May 2018 09:27:38 UTC (273 KB)
[v3] Fri, 18 May 2018 09:40:26 UTC (275 KB)
[v4] Mon, 28 May 2018 12:58:28 UTC (414 KB)
[v5] Wed, 30 May 2018 11:57:48 UTC (414 KB)
[v6] Thu, 31 May 2018 13:04:59 UTC (414 KB)
[v7] Tue, 5 Jun 2018 01:57:20 UTC (415 KB)
[v8] Tue, 2 Oct 2018 12:49:32 UTC (838 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled An Information-Theoretic View for Deep Learning, by Jingwei Zhang and 2 other authors
  • View PDF
  • TeX Source
  • Other Formats
license icon view license
Current browse context:
stat.ML
< prev   |   next >
new | recent | 2018-04
Change to browse by:
cs
cs.LG
stat

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar
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