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
   124   125   126   127   128   129   130   131   132   133   134