Pith. sign in

REVIEW 12 cited by

Towards Understanding the Spectral Bias of Deep Learning

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1912.01198 v3 pith:ACUFIXYO submitted 2019-12-03 cs.LG stat.ML

classification cs.LGstat.ML
keywords neuralbiasspectralnetworkslearningcertaindataexplanation
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

An intriguing phenomenon observed during training neural networks is the spectral bias, which states that neural networks are biased towards learning less complex functions. The priority of learning functions with low complexity might be at the core of explaining generalization ability of neural network, and certain efforts have been made to provide theoretical explanation for spectral bias. However, there is still no satisfying theoretical result justifying the underlying mechanism of spectral bias. In this paper, we give a comprehensive and rigorous explanation for spectral bias and relate it with the neural tangent kernel function proposed in recent work. We prove that the training process of neural networks can be decomposed along different directions defined by the eigenfunctions of the neural tangent kernel, where each direction has its own convergence rate and the rate is determined by the corresponding eigenvalue. We then provide a case study when the input data is uniformly distributed over the unit sphere, and show that lower degree spherical harmonics are easier to be learned by over-parameterized neural networks. Finally, we provide numerical experiments to demonstrate the correctness of our theory. Our experimental results also show that our theory can tolerate certain model misspecification in terms of the input data distribution.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 12 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Neural Spectral Bias and Conformal Correlators I: Introduction and Applications

    hep-th 2026-04 unverdicted novelty 8.0 of 10

    Simple feed-forward neural networks trained on crossing symmetry plus a single anchor value reproduce CFT correlators to percent-level accuracy, and the authors conjecture this works because physical correlators are t...

  2. SONAR: Spectral-Contrastive Audio Residuals for Generalizable Deepfake Detection

    cs.SD 2025-11 conditional novelty 6.0 of 10

    SONAR improves audio deepfake detection by explicitly aligning low- and high-frequency representations for real speech and repelling them for fakes, setting new benchmark EERs on ASVspoof 2021 and in-the-wild data.

  3. Structure-Preserving Patch Decoding for Efficient Neural Video Representation

    cs.CV 2025-06 conditional novelty 6.0 of 10

    Splitting video frames with PixelUnshuffle into structure-preserving patches and decoding them with a global-to-local network improves INR video reconstruction over NeRV-style baselines.

  4. Physics-informed neural network estimation of active material properties in time-dependent cardiac biomechanical models

    cs.LG 2025-05 conditional novelty 6.0 of 10

    A PINN-based method recovers active contractility parameters and detects fibrotic scars in simulated cardiac tissue from sparse displacement and strain data, with a few percent error in homogeneous tests.

  5. Representation Learning for Tabular Data: A Comprehensive Survey

    cs.LG 2025-04 conditional novelty 6.0 of 10

    A comprehensive survey that categorizes deep tabular representation learning into specialized, transferable, and general models, with a feature/sample/objective taxonomy for specialized methods.

  6. SNeRV: Spectra-preserving Neural Representation for Video

    eess.IV 2025-01 conditional novelty 6.0 of 10

    SNeRV decomposes frames with wavelet transforms, embeds only low-frequency content, and regenerates high-frequency details, outperforming prior NeRV models on reconstruction and interpolation.

  7. F$^3$OCUS -- Federated Finetuning of Vision-Language Foundation Models with Optimal Client Layer Updating Strategy via Multi-objective Meta-Heuristics

    cs.CV 2024-11 conditional novelty 6.0 of 10

    F3OCUS combines per-client LNTK layer importance scores with server-side meta-heuristic optimization of layer diversity to improve federated fine-tuning of vision-language models for medical tasks, and releases the 70...

  8. SpiS-GAN: Spiral-Modulated Handwriting Synthesis with Star Operation

    cs.CV 2026-07 conditional novelty 5.0 of 10

    A GAN with elliptical-spiral feature mixing, star-operation blocks, and Sobel edge loss produces more realistic synthetic handwriting and lowers HTR error rates on English and Vietnamese datasets.

  9. FW-GAN: Frequency-Driven Handwriting Synthesis with Wave-Modulated MLP Generator

    cs.CV 2025-08 conditional novelty 5.0 of 10

    FW-GAN combines wave-modulated MLP generation with frequency-domain losses and a wavelet-based discriminator to synthesize handwriting from one example, reporting lower FID scores than prior methods on IAM and HANDS-VNOnDB.

  10. Split Complex-Valued Physics-Informed Neural Networks for Forward and Inverse Nonlinear PDEs

    physics.flu-dyn 2026-07 conditional novelty 4.0 of 10

    Split complex-valued PINNs achieve lower benchmark errors than real-valued PINNs, but the comparison is confounded by doubled parameters and unresolved internal error inconsistencies.

  11. SUPER Module for Detail-Sensitive and Cost-Efficient U-Net Variant Decoders

    cs.CV 2025-11 reject novelty 4.0 of 10

    A plug-in wavelet-domain decoder block improves thin-crack IoU on one self-baseline benchmark, while the abstract's flagship depth-estimation gains and decoder MAC reductions are absent from the main text.

  12. Frequency-Aligned Knowledge Distillation for Lightweight Spatiotemporal Forecasting

    cs.LG 2025-06 reject novelty 4.0 of 10

    The proposed spectral distillation method is not actually evaluated in the experiments, and the abstract's headline error reductions contradict the reported tables.

Pith tools