Skip to content

GAN

Definition

Generative Adversarial Network. A framework with two neural networks, a generator that creates synthetic data and a discriminator that tries to distinguish it from real data. They train together in an adversarial loop until the generator's output is difficult to separate from genuine content.

Full name
Generative Adversarial Network
Components
Generator and discriminator networks
Training method
Adversarial loop
Use in forensics
Basis for many deepfake and synthetic media pipelines

Common questions

How does the adversarial training loop actually improve the generator's output?+

The generator produces synthetic samples and the discriminator scores them as real or fake; the generator is updated to better fool the discriminator while the discriminator is updated to catch the generator's current weaknesses, and repeating this cycle pushes the generator's output progressively closer to the real data distribution.

Why has GAN-based media detection become harder over successive model generations?+

Because the discriminator in training already learns to suppress detectable artifacts in the generator's output, later GAN versions leave fewer of the visual and statistical tells that earlier detection methods relied on, forcing forensic detectors to keep updating to newer artifact classes or model-specific fingerprints.

Related terms

Diffusion Model
A generative neural network architecture (Ho et al., 2020; Stable Diffusion, Rombach et al., 2022) that learns to reverse a noise-addition process...
Encoder-Decoder
A neural architecture where an encoder compresses an input into a compact latent representation and a decoder reconstructs an output image from...
Latent Space
The compressed, lower-dimensional representation of data learned by a neural network's internal layers. Generative models sample from or navigate this space to...
NeRF (Neural Radiance Field)
A neural representation that encodes a 3-D scene as a continuous volumetric function, allowing novel viewpoints to be rendered. In talking-head systems,...
Voice Conversion
A signal-processing or deep-learning technique that transforms the vocal characteristics of a source speaker's utterance to match a target speaker, while preserving...

Explained in

Your journey to becoming a forensic professional starts here.

Practice with mock tests, learn from structured notes, and get your questions answered by a global forensic community, all in one place.