DCGAN: convolutional architecture guidelines for adversarial representations

What changed

Combined strided/fractionally strided convolutions, restricted dense layers, normalization and activation choices to train convolutional adversarial generators and test reusable discriminator features.

Scientific scope

Combined strided/fractionally strided convolutions, restricted dense layers, normalization and activation choices to train convolutional adversarial generators and test reusable discriminator features. Inherits GANs, convolutional networks and batch normalization. Architectural guidelines improve tested training regimes, not a theorem of stability or absence of memorization. Fractionally strided convolution is not a mathematical inverse of ordinary convolution.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits GANs, convolutional networks and batch normalization. Architectural guidelines improve tested training regimes, not a theorem of stability or absence of memorization. Fractionally strided convolution is not a mathematical inverse of ordinary convolution.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction