Page 129 - AI_Ver_3.0_class_11
P. 129
Machine Learning Process
The machine learning process generally involves several key steps to develop and deploy a machine learning model
effectively, as displayed in the following diagram:
1. Preparing data
Machine Learning Process 4. Using features to make and refine predictions
2. Training an algorithm
3. Generating a set of instructions (the model)
on new input data
Features of Machine Learning until the model can accurately make predictions
Following are the features of machine learning:
• It is the science of having machines interpret, analyse and process data as a way to fix real-world problems.
• It learns from data and improve over a period of time. These learnings can be used for automation or prediction.
• It is the dominant mode of AI today.
• It can identify patterns, trends, and relationships within data that may not be immediately apparent to humans.
• It uses data analysis, training, and human review to learn without following specific rules or steps.
Let us now understand the difference between machine learning and deep learning:
Factors Deep Learning Machine Learning
Data Requirement Requires large data Can train on less data
Reliability Provides high accuracy Gives less accuracy
Training Time Takes longer to train Takes less time to train
Hardware dependence Requires GPU to train properly Trains on CPU
Hyperparameter Tuning Can be tuned in various different ways Limited tuning capabilities
Types of Machine Learning
Machine learning is divided into 3 main categories—Supervised, Unsupervised and Reinforcement learning.
Machine Learning
Supervised Learning Unsupervised Learning Reinforcement Learning
Model training with labelled data Model training with unlabelled data Model take actions in the environment
then received state updates and feedbacks
Classification Regression Clustering
Environment
action feedback state
Model
Agent
Let us discuss each in detail.
Introduction: Artificial Intelligence for Everyone 127

