Deep neural networks are a class of machine learning models composed of multiple layers of interconnected nodes (neurons), where each layer transforms its input through learned weights and nonlinear activation functions. The “deep” in DNN refers to the use of many hidden layers—often dozens or hundreds—enabling the model to learn hierarchical and complex representations from raw data.

DNNs are the foundation of most modern AI systems, powering applications in image recognition, natural language processing, speech, and game-playing. Architectures vary depending on the task: convolutional neural networks (CNNs) are common in vision, recurrent or transformer-based networks in language, and feedforward networks in structured data tasks.

Training involves using backpropagation and gradient descent to minimize a loss function, adjusting weights layer by layer. Deeper networks can capture more abstract patterns but require careful regularization (e.g., dropout, batch norm), tuning, and large-scale data to avoid overfitting or vanishing gradients.

While powerful, DNNs are often resource-intensive and opaque, leading to ongoing work in model compression, interpretability, and robustness.