The DOI matches its Crossref record; 12 of the 12 references it cites were found in a registry or an index.
A comparison of deep networks with ReLU activation function and linear spline-type methods
Open accessAbstract
Deep neural networks (DNNs) generate much richer function spaces than shallow networks. Since the function spaces induced by shallow networks have several approximation theoretic drawbacks, this explains, however, not necessarily the success of deep networks. In this article we take another route by comparing the expressive power of DNNs with ReLU activation function to linear spline methods. We show that MARS (multivariate adaptive regression splines) is improper learnable by DNNs in the sense that for any given function that can be expressed as a function in MARS with M parameters there exists a multilayer neural network with O(Mlog(M∕ε)) parameters that approximates this function up to sup-norm error ε. We show a similar result for expansions with respect to the Faber-Schauder system. Based on this, we derive risk comparison inequalities that bound the statistical risk of fitting a neural network by the statistical risk of spline-based methods. This shows that deep networks perform better or only slightly worse than the considered spline methods. We provide a constructive proof for the function approximations.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
The Elements of Statistical Learning
Hastie, T., Friedman, J., & Tibshirani, R. (2001) · Springer Series in Statistics · doi:10.1007/978-0-387-21606-5 · Cited by 24,162
(2010) · Cited by 14,134
Multivariate Adaptive Regression Splines
Friedman, J. H. (1991) · The Annals of Statistics · doi:10.1214/aos/1176347963 · Cited by 8,000
Error bounds for approximations with deep ReLU networks
Yarotsky, D. (2017) · Neural Networks · doi:10.1016/j.neunet.2017.07.002 · Cited by 1,085
Why and when can deep-but not shallow-networks avoid the curse of dimensionality: A review
Poggio, T., Mhaskar, H., Rosasco, L., Miranda, B., & Liao, Q. (2017) · International Journal of Automation and Computing · doi:10.1007/s11633-017-1054-2 · Cited by 559
(2016) · Conference on Learning Theory · Cited by 443
Stochastic gradient descent, weighted sampling, and the randomized Kaczmarz algorithm
Needell, D., Srebro, N., & Ward, R. (2016) · Mathematical Programming · doi:10.1007/s10107-015-0864-7 · Cited by 432
Benefits of Depth in Neural Networks
Telgarsky, M. (2016) · arXiv · doi:10.48550/arxiv.1602.04485 · Cited by 341
Approximation properties of a multilayered feedforward artificial neural network
Mhaskar, H. N. (1993) · Advances in Computational Mathematics · doi:10.1007/bf02070821 · Cited by 157
The Loss Surface of Multilayer Networks.Preprint
Choromanska, A., Henaff, M., Mathieu, M., Arous, G. B., & LeCun, Y. (2014) · arXiv (Cornell University) · Cited by 94
l 1 -regularized neural networks are improperly learnable in polynomial time
Zhang, Y., Lee, J. D., & Jordan, M. I. (2016) · Cited by 90
Limitations of the approximation capabilities of neural networks with one hidden layer
Chui, C. K., Li, X., & Mhaskar, H. N. (1996) · Advances in Computational Mathematics · doi:10.1007/bf02124745 · Cited by 73
Cited by
Activation functions in deep learning: a comprehensive survey and benchmark
Dubey, S. R., Singh, S. K., & Chaudhuri, B. B. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.06.111 · Cited by 1,259
A survey on modern trainable activation functions
Apicella, A., Donnarumma, F., Isgrò, F., & Prevete, R. (2021) · Neural Networks · doi:10.1016/j.neunet.2021.01.026 · Cited by 571
Wang, Y., Li, Y., Song, Y., & Rong, X. (2020) · Applied Sciences · doi:10.3390/app10051897 · Cited by 320
Toğaçar, M., Ergen, B., & Cömert, Z. (2020) · Medical Hypotheses · doi:10.1016/j.mehy.2019.109531 · Cited by 310
Qin, B., & Li, D. (2020) · Sensors · doi:10.3390/s20185236 · Cited by 296
Metadata
- Authors
- Eckle, K., & Schmidt-Hieber, J.
- Venue
- Neural Networks
- Publisher
- Elsevier BV
- Published
- 2019-02
- Volume
- 110
- Pages
- 232-242
- ISSN
- 0893-6080 (ISSN-L)
- OpenAlex
- W2796199633
- Cited by
- 409
- References
- 12
- FWCI
- 16.83
- Type
- journal article
- Language
- en
- Open access
- hybrid
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.