Page 23 - CT_AI_Class-6
P. 23

Ask
                                                                      AGENT
                                                       OrangeAI
                  If AI can only learn from data, how can it be sure it's making the right choice?

                          Study



                 HOW DOES AI LEARN?

                 Artificial Intelligence learns by collecting data, learning from it, making decisions and improving
                 over time through past experiences. Data can be classified as labelled data and unlabelled
                 data.

                   Labelled data: This is data that has been given a clear tag or label. Each piece of data has the
                    correct answer or category written alongside it.
                   Unlabelled data: This is data without any tags or labels. The data doesn't tell the AI what it is.

                    The AI has to figure out the patterns and similarities on its own.
                 Let’s break down each step of how AI learns:

                   Data collection: AI systems operate on vast datasets, making data collection crucial. They
                    gather information from various sources such as sensors, user activity, websites, cameras
                    and applications. This data can be unstructured, like images, speech or videos or structured,
                    consisting of numbers in a table. For example, a fitness app can collect data from your wearable
                    devices, such as a smartwatch, regarding the number of steps, heart rate and sleep data.

                   Learning from data: AI learns from the data provided to it. During the model training phase,
                    the AI model adjusts itself based on data to learn how to perform tasks and make predictions.
                    It also finds patterns in the data. For example, AI learns to recognise cats in photos by analysing
                    thousands of images labelled as “cat.”

                   Decision making: After learning from the data, the AI makes decisions or predictions. These
                    decisions  are based  on the  patterns  it  detected during the  training  phase.  For example,
                    a recommendation  system  decides  which film to  show you  next  based  on your  past
                    viewing habits.
                   Improvement: AI systems improve over time by receiving feedback and analysing results.

                    They retrain or update their models using new data or corrections. For example, a self-
                    driving car improves its driving decisions as it encounters more traffic situations and receives
                    updates.



                                   What is the next number in the sequence: 17, 35, 52, 69, __?

                                           a)  85                                             b)  90

                                           c)  86                                             d)  95








                                                                           Introduction to AI & Everyday Examples  21
   18   19   20   21   22   23   24   25   26   27   28