Pith. sign in

REVIEW 8 cited by

On the Relationship between Self-Attention and Convolutional Layers

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 1911.03584 v2 pith:5HTKUZHR submitted 2019-11-08 cs.LG cs.CLcs.CVstat.ML

classification cs.LGcs.CLcs.CVstat.ML
keywords layersattentionconvolutionalself-attentionconvolutionlayersimilarlyvision
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Recent trends of incorporating attention mechanisms in vision have led researchers to reconsider the supremacy of convolutional layers as a primary building block. Beyond helping CNNs to handle long-range dependencies, Ramachandran et al. (2019) showed that attention can completely replace convolution and achieve state-of-the-art performance on vision tasks. This raises the question: do learned attention layers operate similarly to convolutional layers? This work provides evidence that attention layers can perform convolution and, indeed, they often learn to do so in practice. Specifically, we prove that a multi-head self-attention layer with sufficient number of heads is at least as expressive as any convolutional layer. Our numerical experiments then show that self-attention layers attend to pixel-grid patterns similarly to CNN layers, corroborating our analysis. Our code is publicly available.

Discussion (0). Sign in to comment.

Forward citations

Cited by 8 Pith papers

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

  1. RoVE: Rotary Value Embeddings Attention for Relative Position-dependent Value Pathways

    cs.LG 2026-06 unverdicted novelty 6.0 of 10

    RoVE rotates value embeddings simultaneously with keys in attention to make values position-dependent, reframing RoPE as attentive convolution and reporting gains on long-context tasks in 124M and 354M GPT-2 models.

  2. Long-Term Visual Localization in Dynamic Benthic Environments: A Dataset, Footprint-Based Ground Truth, and Visual Place Recognition Benchmark

    cs.CV 2026-03 conditional novelty 6.0 of 10

    A benchmark shows state-of-the-art visual place recognition performs poorly on a new multi-site, multi-year benthic AUV dataset, and that distance-based ground truth inflates recall.

  3. Numerical Investigation of Sequence Modeling Theory using Controllable Memory Functions

    cs.LG 2025-06 conditional novelty 6.0 of 10

    A synthetic benchmark with controllable memory functions maps how four sequence architectures handle exponential, polynomial, sparse-spike, and Airy temporal structures.

  4. GASPnet: Global Agreement to Synchronize Phases

    cs.LG 2025-07 conditional novelty 5.0 of 10

    A CNN augmented with global attention-driven phase synchronization (GASPnet) outperforms a parameter-matched CNN on noisy multi-object and superimposed-image classification.

  5. Low-latency vision transformers via large-scale multi-head attention

    cs.CV 2025-06 conditional novelty 5.0 of 10

    Attention heads in compact vision transformers each recognize small label subsets with little noise, which the authors exploit for diverse ensembles and low-latency hybrid architectures on CIFAR-100.

  6. Efficient Feedback Gate Network for Hyperspectral Image Super-Resolution

    eess.IV 2025-06 conditional novelty 5.0 of 10

    A group-based deep network (EFGN) with channel-shuffle/dilated feedback and large-kernel gating achieves state-of-the-art hyperspectral image super-resolution on Chikusei, Pavia Centre, and Harvard.

  7. Compute Can't Handle the Truth: Why Communication Tax Prioritizes Memory and Interconnects in Modern AI Infrastructure

    cs.DC 2025-07 reject novelty 4.0 of 10

    A CXL-based disaggregated memory architecture with hybrid XLink interconnects is proposed and prototyped, claiming large speedups for memory-bound AI and HPC workloads.

  8. Graph Collaborative Attention Network for Link Prediction in Knowledge Graphs

    cs.LG 2025-07 reject novelty 2.0 of 10

    GCAT is presented as a new graph attention model for knowledge graph link prediction, but its equations are those of KBGAT and its reported benchmark numbers do not support the stated performance claims.

Pith tools