Generative adversarial network
A generative adversarial network is a generative model with a generator and discriminator trained in competition, enabling the generator to produce samples resembling a target dataset or distribution.
Inside Deep Learning: Math, Algorithms, Models
A generative deep-learning model consisting of a generator and discriminator trained adversarially to produce realistic samples.
- How it is used
- Uses adversarial training to create realistic synthetic data and manipulate generated content for one-to-many problems.
- How deep it goes
- Explains G, D, latent vectors, losses, alternating PyTorch updates, mode collapse, WGANs, and latent manipulation, but skips Wasserstein derivation.
- Examples
- Demonstrates MNIST digit generation, generated-person images, latent interpolation, semantic image edits, loss curves, and discriminator confidence scores.
- What you can do after
- Builds a fully connected PyTorch GAN, alternates discriminator and generator updates, visualizes outputs, and interprets latent-space manipulations.
Read how Inside Deep Learning: Math, Algorithms, Models explains generative adversarial network →
Learn Generative AI with PyTorch
A generative adversarial network, a model family trained with a generator and discriminator to synthesize realistic data or transform data between domains.
- How it is used
- Uses competing generator and discriminator networks to create samples resembling training data or transform image attributes.
- How deep it goes
- Develops a build-train-stop-use workflow, including data preparation and early stopping in a number-pattern project.
- Examples
- Works through exponential-curve data and sequences of multiples of 5, alongside face-image transformation and music generation.
- What you can do after
- Build it from scratch in PyTorch, then train, save, load, and use a generator to produce patterned samples.
Read how Learn Generative AI with PyTorch explains generative adversarial network →
Outlier Detection in Python
A generative adversarial network is a model consisting of a generator and discriminator trained competitively so the generator produces data resembling a reference distribution.
- How it is used
- Uses GANs as generative outlier detectors for tabular and image data.
- How deep it goes
- Explains joint training, latent input, and discriminator and BiGAN scoring, without detailing loss functions.
- Examples
- Shows generator–discriminator diagrams, BiGAN prediction paths, and a PyOD ALAD fit-and-score listing.
- What you can do after
- Apply it via PyOD’s ALAD detector to fit data, inspect decision_scores_, and plot learning curves.
Read how Outlier Detection in Python explains generative adversarial network →
Grokking Machine Learning
A generative adversarial network is a generative model composed of competing generator and discriminator networks that learn to produce realistic data.
- How it is used
- Uses GANs to illustrate how neural networks can create images rather than merely make predictions.
- How deep it goes
- Defines generator and discriminator roles and their training interaction, but leaves generative learning largely unexplored.
- Examples
- Illustrates GANs with face-image generation and names paintings, music, and stories as generative outputs.
- What you can do after
- Understand how competing networks can learn to produce realistic-looking images from a dataset.
Read how Grokking Machine Learning explains generative adversarial network →
Sutskever's List
A generative adversarial network is a model architecture in which a generator creates candidates and a discriminator learns to distinguish them from genuine data.
- How it is used
- Frames GANs as a mechanism for producing convincing synthetic media and as a source of dual-use safety concerns.
- How deep it goes
- Defines the generator-discriminator competition through analogy but stops short of explaining training objectives or implementation details.
- Examples
- Uses a counterfeit-artwork analogy and realistic synthetic videos or images as examples.
- What you can do after
- Evaluate how adversarial generation can produce convincing forgeries and create risks around authenticity and misuse.
Read how Sutskever's List explains generative adversarial network →
Generative AI in Action
A generative adversarial network is a generative model composed of a generator and discriminator trained adversarially to produce realistic synthetic data.
- How it is used
- Selects GANs for complex, diverse data when image realism matters, including generation, style transfer, super-resolution, and augmentation.
- How deep it goes
- Explains generator–discriminator competition, prompt conditioning, realism scoring, gradient updates, and latent space without mathematical derivation.
- Examples
- Uses realistic image generation, artistic style transfer, chair design, and medical or space-image super-resolution.
- What you can do after
- Evaluate GAN suitability for high-realism image generation and explain its adversarial training workflow.
Read how Generative AI in Action explains generative adversarial network →
Evolutionary Deep Learning: Genetic algorithms and neural networks
A generative adversarial network is a model consisting of a generator and discriminator trained in opposition to produce realistic synthetic data.
- How it is used
- Uses adversarial training to generate realistic synthetic data and explore evolutionary optimization of difficult GAN training.
- How deep it goes
- Explains network roles, feedback, implementation, artifacts, and experiments, but stops short of extensive mathematical derivation.
- Examples
- Demonstrates art forgery, class-specific handwritten digits, face classification, and experiments with GAN hyperparameters and datasets.
- What you can do after
- Build it from scratch with Python and Keras, inspect training failures, and experiment with evolutionary hyperparameter optimization.
Data Without Labels
A generative adversarial network is a generative model consisting of a generator and discriminator trained adversarially to produce realistic samples.
- How it is used
- Supports synthetic-data generation, image creation, style transfer, and data augmentation across business and machine-learning applications.
- How deep it goes
- Explains the generator, discriminator, adversarial training, applications, and major variants without detailing optimization mathematics or implementation code.
- Examples
- Uses image generation, class-controlled synthesis, style transfer, image translation, high-resolution imagery, and dataset expansion as examples.
- What you can do after
- Evaluate GAN variants for synthetic-data generation, image creation, style transfer, and image-to-image translation.
Read how Data Without Labels explains generative adversarial network →
Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI
A generative adversarial network is a deep learning architecture consisting of a generator and discriminator trained adversarially to produce samples resembling real data.
- How it is used
- Supports synthetic tabular data generation for scarcity, diversity, edge-case testing, and privacy-conscious experimentation.
- How deep it goes
- Defines generator–discriminator interaction and random-noise generation but stops short of mathematical or implementation details.
- Examples
- Uses tabular scarcity, minority classes, edge cases, privacy-preserving fictitious records, and German credit data generation.
- What you can do after
- Evaluate/understand how adversarial networks generate synthetic records and how synthetic data can be tested.
Introduction to Generative AI, Second Edition
A generative adversarial network is a model composed of a generator and discriminator trained through competition between synthetic-data creation and authenticity evaluation.
- How it is used
- Creates or alters synthetic images and videos through competition between generation and authenticity evaluation.
- How deep it goes
- Explains the generator–discriminator idea and historical position, but remains an overview rather than an implementation treatment.
- Examples
- Uses image and video creation or alteration as the main application context, alongside comparisons with autoencoders and diffusion models.
- What you can do after
- Evaluate GANs as a synthetic-media technique and distinguish their historical role from diffusion models.
Read how Introduction to Generative AI, Second Edition explains generative adversarial network →
Grokking AI Algorithms, Second Edition
A generative adversarial network is a machine-learning model consisting of a generator and a discriminator trained through competition to produce increasingly realistic synthetic data.
- How it is used
- Uses competing networks to generate realistic synthetic images, videos, and three-dimensional objects.
- How deep it goes
- Explains the generator–discriminator loop and feedback intuition, but stops short of implementation or mathematical training details.
- Examples
- Uses landscapes, deepfakes, hairstyle overlays, and 3D chairs generated from 2D images.
- What you can do after
- Understand how generator–discriminator competition improves synthetic data and supports image-generation applications.
Read how Grokking AI Algorithms, Second Edition explains generative adversarial network →
Build a Text-to-Image Generator (from Scratch)
A generative adversarial network is a generative model trained through competition between a generator that creates samples and a discriminator that distinguishes generated samples from real data.
- How it is used
- Explains adversarial image generation and supplies the generative component underlying VQGAN.
- How deep it goes
- Defines the two-network minimax process and its ideal outcome, while deferring implementation and detailed loss equations.
- Examples
- Uses random noise, real and synthetic samples, image synthesis, music generation, and VQGAN image-token conversion as examples.
- What you can do after
- Evaluate GAN generation through generator-discriminator behavior, adversarial training dynamics, and known failure modes.
Read how Build a Text-to-Image Generator (from Scratch) explains generative adversarial network →
Related concepts
- generatorhas part · A generator is a computational component or construct that produces data or values from inputs, representations, context, or specified constraints.
- discriminatorhas part · A discriminator is the model component of a generative adversarial network that distinguishes real data from data generated by its generator.
- VQGANincludes · VQGAN is a generative model that combines vector-quantized discrete latent representations with adversarial and perceptual objectives to synthesize images.
- generative modelingkind of · Generative modeling is a machine-learning approach that learns patterns or a data distribution to generate new samples resembling the data.
- DCGAN / deep convolutional ganincludes · A deep convolutional generative adversarial network is a GAN whose generator and discriminator use deep convolutional neural-network architectures.
- generationkind of · Generation is the computational production of outputs from supplied inputs or context, often by applying a model to produce predictions or new content.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.