GAN (Generative Adversarial Network)
A generative model architecture consisting of two neural networks, a generator and a discriminator, that compete against each other, with the generator learning to create increasingly realistic outputs.
GANs train through an adversarial game. The generator creates fake samples (images, text, audio), and the discriminator tries to distinguish fake from real. As training progresses, the generator improves at creating convincing fakes while the discriminator improves at detecting them. The equilibrium produces a generator that creates highly realistic outputs.
GANs dominated image generation before diffusion models and produced breakthroughs like StyleGAN (photorealistic face generation), Pix2Pix (image-to-image translation), and CycleGAN (unpaired style transfer). Their key advantage is the ability to generate sharp, high-resolution images. However, they suffer from training instability, mode collapse (generating limited variety), and difficulty evaluating output quality.
While diffusion models have surpassed GANs in most image generation benchmarks, GANs remain relevant for specific applications: real-time generation (GANs are much faster than diffusion models at inference), image-to-image translation tasks, and data augmentation for training other models. For growth teams, GAN-based techniques can generate synthetic training data, create product variations for A/B testing visual designs, and power style transfer features in consumer applications.
Related Terms
RAG (Retrieval-Augmented Generation)
A technique that grounds LLM responses in external data by retrieving relevant documents at query time and injecting them into the prompt context.
Embeddings
Dense vector representations of text, images, or other data that capture semantic meaning in a high-dimensional space, enabling similarity search and clustering.
Vector Database
A specialized database optimized for storing, indexing, and querying high-dimensional vector embeddings with sub-millisecond similarity search.
LLM (Large Language Model)
A neural network trained on massive text corpora that can generate, understand, and transform natural language for tasks like summarization, classification, and conversation.
Fine-Tuning
The process of further training a pre-trained LLM on a domain-specific dataset to specialize its behavior, style, or knowledge for a particular task.
Prompt Engineering
The practice of designing and iterating on LLM input instructions to reliably produce desired outputs for a specific task.