renzume on Nostr: FFTNet introduces a novel approach to sequence processing using Fast Fourier ...
FFTNet introduces a novel approach to sequence processing using Fast Fourier Transform, achieving O(n log n) complexity compared to traditional self-attention's quadratic complexity. The framework employs spectral filtering and modReLU activation to efficiently capture long-range dependencies, demonstrating superior performance on Long Range Arena and ImageNet benchmarks.
https://arxiv.org/abs/2502.18394#machinelearning #neuralnetworks #fft #self-attention #performanceoptimization
Published at
2025-02-26 13:18:09Event JSON
{
"id": "c0eb1eb81634df11025b701bcf09b88964786b06ef7a8f0a0778c157facf952b",
"pubkey": "d3972a5c762e9cab61c5404c2f673480022b90860ead779d3f5eef5cbe7a7640",
"created_at": 1740575889,
"kind": 1,
"tags": [],
"content": "FFTNet introduces a novel approach to sequence processing using Fast Fourier Transform, achieving O(n log n) complexity compared to traditional self-attention's quadratic complexity. The framework employs spectral filtering and modReLU activation to efficiently capture long-range dependencies, demonstrating superior performance on Long Range Arena and ImageNet benchmarks.\nhttps://arxiv.org/abs/2502.18394\n#machinelearning #neuralnetworks #fft #self-attention #performanceoptimization",
"sig": "533e88c6ff42e04ddd21d343fd79038233d97754c51b6e0e520c02edafe91fe99888a134a3c7cb08761aeab8eb83bf5dcf13f2514466a770f0f3e0287c32d3d9"
}
