Page 138 - Artificial Intellegence_v2.0_Class_12
P. 138

Unsupervised Learning Algorithm


         K Means Clustering
                                          efore    eans                  After    eans






                                                         eans







          his unsupervised learning algorithm divides data sets into a predetermined number of clusters in a  ay that makes
         each cluster's data homogeneous and distinct from the data in the other clusters.


         Other examples include – Anomaly detection, Neural Networks


                                            Reinforcement Learning Algorithm
          Q-Learning

                                                     nvironment












                                        Action                           tate
                                                         Agent          e ard


           his is a model free, off policy    algorithm and is based on the  ellman  quation.  he technique makes use of
          a   table, a lookup database that records the agent's estimated utility or  quality  of doing a specific action in a
          specific condition.  hrough trial and error, the agent updates the   values to ma imise it in the table and eventually
          converges to the best possible policy.
          Other examples include – SARSA (State–action–reward–state–action), DQN (Deep Q Network)





                Showcase through a Compelling Story

         he follo ing story illustrates through a series of visuals ho  a pro ect made a      difference in the lives of thousands
        of Indian  omen.







              136     Touchpad Artificial Intelligence (Ver. 2.0)-XII
   133   134   135   136   137   138   139   140   141   142   143