Which of the following is an example of unsupervised neural network?

Answer: Self-organizing feature map

Explanation: A self-organizing feature map (SOM) is a prototypical unsupervised neural network. SOMs learn patterns and cluster input data by adjusting weights based only on input samples, without any labeled target outputs, and they produce a topologically organized map of the input space.

• Back-propagation network: Trained with labeled input–output pairs using an error signal and gradient descent, so it is a supervised learning method.

• Hebb network: Uses the Hebbian learning rule, which is an unsupervised weight update based on input correlations. While Hebbian learning is unsupervised, the question asks for the standard architecture that exemplifies unsupervised clustering and topology preservation—SOM is the canonical example.

• Associative memory network: Designed for storing and recalling patterns (e.g., Hopfield networks). Their learning may be unsupervised in some formulations, but they are mainly used for pattern retrieval rather than general clustering or topology-preserving mapping.

• Self-organizing feature map: Learns to cluster inputs and form a topographic map without labeled targets, making it the correct and most representative example of an unsupervised neural network.