Dual-Perspective Activation: Efficient Channel Denoising via Joint Forward-Backward Criterion for Artificial Neural Networks

Tian Qiu, Chenchao Gao, Zunlei Feng, Jie Lei, Bingde Hu +3 more
2/3/2026

No official code found. Best implementation: horrible-dong/DPA (8 stars, Apache-2.0, last commit 2025-11-26).

Abstract

.

DOISemantic Scholar

Code Implementations(1)

horrible-dong/DPACommunity74%
Apache-2.0

[NeurIPS 2024] Dual-Perspective Activation: Efficient Channel Denoising via Joint Forward-Backward Criterion for Artificial Neural Networks

81PythonOct 30, 202410 months agoApache-2.0
activation-functionartificial-neural-networkschannel-denosingdeep-learningmachine-learning+3 more

Cite this paper

@article{qiu2026dualperspective,
  title  = {Dual-Perspective Activation: Efficient Channel Denoising via Joint Forward-Backward Criterion for Artificial Neural Networks},
  author = {Tian Qiu and Chenchao Gao and Zunlei Feng and Jie Lei and Bingde Hu and Xingen Wang and Yi Gao and Mingli Song},
  year   = {2026},
  doi    = {10.52202/079017-2254},
  url    = {https://doi.org/10.52202/079017-2254},
  journal = {NEURIPS 2024 2024}
}

Discussion