1. Supervised Learning:
In supervised learning, a model is trained with a dataset containing both input and output data. Each example in the dataset consists of an input feature set (e.g., images, text, or numerical data) and the corresponding output label (e.g., categories, numbers, or classifications). The goal of the model is to learn a function that maps input data to the correct output data. Once the model is trained, it can be used to make predictions for new, unknown data. Supervised learning is the most commonly used method in machine learning and is employed in applications such as image recognition, speech recognition, and spam detection.

2. Unsupervised Learning
In unsupervised learning, models are trained with datasets that do not contain output data. Instead, the model attempts to identify and learn patterns or structures in the input data itself. There are various approaches in unsupervised learning, including clustering (grouping similar data points), dimensionality reduction (reducing the number of features to better visualize or process data), and anomaly detection (identifying data points that deviate from the norm). Unsupervised learning is often used in applications such as customer segmentation, recommendation systems, and text mining.

3. Semi-supervised Learning
Semi-supervised learning is a method that falls between supervised and unsupervised learning. In this approach, a model is trained using a dataset that contains both labeled (input-output pairs) and unlabeled data (input data without corresponding output labels). Typically, the amount of labeled data is limited, while the volume of unlabeled data is much larger.
The goal of semi-supervised learning is to leverage the information contained in the unlabeled data to improve the model’s performance on labeled data. This can be achieved through various techniques, such as clustering, self-training, or graph-based methods. By using the structure and patterns found in the unlabeled data, the model can make better predictions when presented with new labeled data.
Semi-supervised learning is useful in situations where obtaining labeled data is expensive, time-consuming, or requires expert knowledge. It has been applied in various domains, including image and speech recognition, natural language processing, and bioinformatics.

4. Reinforcement Learning
Reinforcement learning is a method of machine learning in which an agent interacts with an environment and aims to achieve a goal by performing actions and receiving rewards or penalties. The agent’s goal is to learn an optimal strategy (also referred to as a „policy“) that maximizes cumulative reward over time. Reinforcement learning differs from supervised and unsupervised learning in that it is based on interactions with the environment and does not use pre-defined examples for learning. Applications of reinforcement learning include games (e.g., chess, Go), robotics, autonomous driving, and resource allocation.