Unsupervised Learning

Unsupervised Learning is a paradigm of Machine Learning in which an Algorithm discovers patterns and structure in a Dataset without externally supplied Labels or target outputs. Unlike Supervised Learning, it identifies latent relationships, often by optimizing an Objective Function based on properties of the input data rather than labeled input–output pairs. Common approaches include K-Means Clustering, Principal Component Analysis, and unsupervised Generative Modeling. Software support includes the Scikit-learn library and general-purpose Machine Learning and Deep Learning frameworks such as PyTorch and TensorFlow; these tools are not specific to unsupervised learning. Self-Supervised Learning, often considered a form of unsupervised learning, constructs training targets from the data itself and is widely used in training Large Language Models. Applications of unsupervised learning include discovering Clusters, performing Anomaly Detection, and obtaining lower-dimensional representations through Dimensionality Reduction, particularly in high-dimensional Data Science workflows.