Pith. sign in

REVIEW 7 cited by

PolySketchFormer: Fast Transformers via Sketching Polynomial Kernels

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 2310.01655 v3 pith:PSKI4TU4 submitted 2023-10-02 cs.LG

classification cs.LG
keywords attentionpolynomiallanguagemodelspolysketchformertrainingachievesapproximation
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The quadratic time and memory complexity inherent to self-attention mechanisms, with respect to sequence length, presents a critical computational bottleneck in the training and deployment of large-scale Transformer-based language models. Recent theoretical results indicate the intractability of sub-quadratic softmax attention approximation under reasonable complexity assumptions. This paper addresses this challenge by first demonstrating that polynomial attention with high degree can effectively replace softmax without sacrificing model quality. Next, we develop polynomial sketching techniques from numerical linear algebra to achieve linear-time polynomial attention with approximation guarantees. Crucially, our approach achieves this speedup without requiring the sparsification of attention matrices. We also present a block-based algorithm to apply causal masking efficiently. Combining these techniques, we provide \emph{PolySketchFormer}, a practical linear-time Transformer architecture for language modeling that offers provable guarantees. We validate PolySketchFormer empirically by training language models capable of handling long contexts. These experiments utilize both synthetic and real-world datasets (PG19, Wikipedia and C4) on Google Cloud TPUs. For context lengths of 32k and GPT-2 style models, our model achieves a 2.5-4x speedup in training compared to FlashAttention, with no observed degradation in quality across our experiments.

Discussion (0). Sign in to comment.

Forward citations

Cited by 7 Pith papers

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

  1. Selective Rotary Position Embedding

    cs.CL 2025-11 unverdicted novelty 7.0 of 10

    Selective RoPE replaces RoPE's fixed rotation angles with input-dependent, learnable angles and improves recall-focused tasks in gated linear and softmax transformers.

  2. Norming Sets for Tensor and Polynomial Sketching

    math.NA 2025-06 conditional novelty 7.0 of 10

    Norming sets are used to bound sketching dimensions for algebraic varieties and polynomial images under arbitrary sketch operators, including a new median sketch that needs only about dim(V) structured measurements.

  3. ELiTeFormer: An Efficient Transformer for FPGAs

    cs.AR 2026-07 conditional novelty 6.0 of 10

    Hybrid linear attention plus ternary projections, co-designed with a multiplier-free PE, deliver 10× weight and 12.8× KV-cache compression with competitive MMLU and FPGA speedups over LLaMA 3 on A100.

  4. A Theoretical Study of (Hyper) Self-Attention through the Lens of Interactions: Representation, Training, Generalization

    cs.LG 2025-06 conditional novelty 5.0 of 10

    Single-layer linear self-attention can represent, train on, and length-generalize pairwise interaction functions under data-versatility and exact-realizability assumptions, and the paper introduces higher-order HyperA...

  5. Neural Interpretable PDEs: Harmonizing Fourier Insights with Attention for Scalable and Interpretable Physics Discovery

    cs.LG 2025-05 conditional novelty 5.0 of 10

    NIPS is a neural operator that uses linear attention and Fourier kernels to simultaneously predict PDE solutions and recover hidden material properties from limited data.

  6. Unifying Learning Dynamics and Generalization in Transformers Scaling Law

    cs.LG 2025-12 reject novelty 4.0 of 10

    Claims a two-stage transformer scaling law (exponential then C^{-1/6}) with matching bounds, but the lower bounds are missing, the exponent is inconsistent (-1/7 vs -1/6), and the law is an artifact of hand-set M = Θ(...

  7. Scaling Context Requires Rethinking Attention

    cs.LG 2025-07 conditional novelty 4.0 of 10

    On 64k-token natural language training, power attention with degree 2 achieves lower loss per FLOP than both softmax attention and existing linear attention, and its GPU kernels run faster than Flash Attention at long...

Pith tools