REVISITING AI, ML, AND DL

• Artificial Intelligence (AI): The overarching field of study that enables machines to mimic human intelligence. AI systems work based on algorithms and data to produce a desired output.
• Machine Learning (ML): A subset of AI where machines learn from experience and improve on tasks with new data. ML models learn from mistakes and self-correct for future executions.
• Deep Learning (DL): A very specific learning approach and a subset of Machine Learning. DL enables software to train itself with vast amounts of data and can develop its own algorithms.
• Relationship: AI is the umbrella term, with ML as a subset, and DL as a further subset of ML.

TYPES OF AI MODELS

AI models are broadly classified into two categories: Rule-based and Learning-based.

• Rule-Based Approach: The developer explicitly defines the rules and patterns for the machine to follow. The learning is static, and the model does not adapt to changes in the data or new exceptions. An example is a chatbot with a predefined set of questions and answers.
• Learning-Based Approach: The machine learns on its own by analyzing examples and finding patterns or rules from data. These models are adaptive and can modify themselves to handle changes in data and exceptions. A spam email filter is a classic example.

MACHINE LEARNING APPROACHES

Learning-based approaches are further divided into three types:

1. Supervised Learning

This model uses labeled data for training, where the dataset is tagged with information that acts as a guide. The goal is to predict a specific output based on the input data, similar to a teacher-student relationship.

• Examples: Predicting the currency of a coin based on its weight, or classifying images as "dog" or "cat".

Sub-categories:

• Classification: Predicts a discrete value or a class label. Examples include classifying emails as spam or not spam and predicting whether a customer is eligible for a bank loan (yes/no).
• Regression: Predicts a continuous value. Examples include predicting house prices, stock prices, or tomorrow's temperature.

2. Unsupervised Learning

This model works on unlabeled data. The purpose is to identify hidden relationships, patterns, and trends in the data without any human guidance. The model groups data points based on similarities.

• Examples: A supermarket clustering customers into groups based on purchasing habits, or an OTT platform recommending movies based on watch history.

Sub-categories:

• Clustering: Groups data points into clusters based on similarities.
• Association: Finds meaningful relationships between variables in a dataset. An example is recommending butter to a customer who buys bread.

3. Reinforcement Learning

This approach involves a computer making a series of decisions to maximize a "reward". It learns through a trial-and-error process with positive and negative feedback. It is useful for complex problems where pre-existing data is insufficient.

• Examples: A self-driving car learning to park or a humanoid learning to walk.

SUMMARY OF ML MODELS

• Supervised learning models are used when we want to determine relationships through training.
• Unsupervised learning models are used when we want to discover new patterns from data.
• Reinforcement learning models are used when we want to implement machine learning through a reward mechanism.

SUB-CATEGORIES OF SUPERVISED LEARNING MODEL

Classification Model

Here the data is classified according to the labels. For example, in the grading system, students are classified on the basis of the grades they obtain with respect to their marks in the examination. This model works on discrete dataset which means the data need not be continuous.

Regression Model

Regression: Such models work on continuous data. For example, if you wish to predict your next salary, then you would put in the data of your previous salary, any increments, etc., and would train the model. Here, the data which has been fed to the machine is continuous.

SUB-CATEGORIES OF UNSUPERVISED LEARNING MODEL

Clustering Model

A clustering model is a type of unsupervised learning that works on unlabeled data to find hidden patterns and group similar data points together. The model's main purpose is to independently organize data into "clusters" based on similarities, helping users discover the underlying structure of their data without any prior guidance.

Association Model

An association model is a type of unsupervised learning that discovers interesting relationships and co-occurrences among variables in a dataset. It is often used for "market basket analysis," which finds items that are frequently bought together.

DEEP LEARNING AND NEURAL NETWORKS

Artificial Neural networks (ANN) - Artificial Neural networks are modelled on the human brain and nervous system. They are able to automatically extract features without input from the programmer. Every neural network node is essentially a machine learning algorithm. It is useful when solving problems for which the data set is very large.

Convolutional Neural Network (CNN) - Convolutional Neural Network is a Deep Learning algorithm which can take in an input image, assign importance (learnable weights and biases) to various aspects/objects in the image and be able to differentiate one from the other.

NEURAL NETWORK

Neural networks are loosely modelled after how neurons in the human brain behave. The key advantage of neural networks is that they are able to extract data features automatically without needing the input of the programmer. A neural network is essentially a system of organizing machine learning algorithms to perform certain tasks.

It is a fast and efficient way to solve problems for which the dataset is very large, such as in images.

STRUCTURE OF A NEURAL NETWORK

A Neural Network is divided into multiple layers and each layer is further divided into several blocks called nodes. Each node has its own task to accomplish which is then passed to the next layer.

• Input Layer: Acquires data and feeds it to the network without processing it.
• Hidden Layers: The core processing happens here. Each node in this layer performs its own algorithm. Information is processed using weights and biases.
• Output Layer: Receives the final processed data and presents it as the final output to the user.