Page 29 - Ai Robogenius
P. 29

EVALUATION—THE TESTING STAGE

                Evaluation is one of the most critical stages of an AI project. This phase is an important step
                where we check how well the AI model works to see if it meets the goals set at the start. If the

                model does not perform as expected, changes may be needed. Once the developer is happy
                that the model works well and fits the project’s needs, the AI project moves to the next phase
                called deployment.



                               Task                                                          21 st
                                                                                            Century
                                                                                             Skills  #Critical Thinking
                   Case: Substantial Electricity Bill (Continued…)
                   After installing the sensor and alarm system in some classrooms, gather the following data:

                   Reduction in electricity usage compared to the period before the system was installed
                   Number of times the alarm went off and what actions were taken

                   Feedback from teachers and students on how effective and easy to use the system is

                   A comparison of the costs of installing and maintaining the system versus the money saved
                   on electricity
                   Any technical problems or false alarms during the trial period

                   Use this data to decide if the system effectively reduces electricity usage and if it can be used
                   in all classrooms. Use this evaluation to improve the system before fully implementing it.




                DEPLOYMENT

                In this phase, the best model is put into real use. It is added to the actual system where people
                can use it in everyday work. A system is also set up to watch the model’s performance regularly,
                making sure it keeps giving correct and reliable results. If problems come up, like changes in data
                or lower performance, the model is fixed and updated. The goal is to keep the model working well
                over time and make improvements as needed.


                               Task                                                          21 st
                                                                                            Century
                                                                                            Skills  #Information Literacy
                   Case: Substantial Electricity Bill (Continued…)

                   Start by installing the solution (the app and sensors) in a few classrooms. After testing and
                   reviewing the results, including the cost of installation and the reduction in electricity use,
                   check if the system helps reduce the bill. If it does, roll out the system to all classrooms. Collect
                   feedback to improve the system based on how it is used in real-time.







                                                                                 Stages of AI Project Cycle
                                                                                                                    27
   24   25   26   27   28   29   30   31   32   33   34