Page 165 - Artificial Intellegence_v2.0_Class_12
P. 165

B.   Fill in the blanks.
                1.   Human biases in selecting test data can negatively impact      .

                2.                  is an example of detrimental effects that can occur over time on the AI model.

                3.   For a balanced dataset,           may be a useful choice for evaluating a classification model.
                4.   The               is used to assist us in identifying the important factors associated with the problem.

                5.   The initial insights gained help to get an understanding of the data and later on, help in algorithm selection, metrics
                     choice, etc. This complete procedure is called       .
                6.   The               constitutes the largest proportion of the original dataset and is employed to train the machine
                     learning models.

                7.   The               is a subset of data employed for unbiased evaluation of model fit and hyperparameter tuning.
            C    State  hether the  ollo ing statements are true or  alse
                1.   If the data you collect is good, you will be able to build an effective AI algorithm, and your whole
                     project will collapse.
                2.   Human entries are always prone to mistakes.
                3.   Building the algorithm corresponds to Evaluation stage of the AI model cycle.

                4.   The AI model gets trained on the data fed to it and then is able to design a model which is
                     adaptive to the change in data.
                5.   Various AI development platforms provide substantial documentation to assist development teams.
            D    Match the  ollo ing

                1.   Training set                     a.   4Ws Problem Canvas
                2.   BigML                            b.   data cleaning
                3.   Data Exploration                 c.   AI development Platform
                4.   Problem Scoping                  d.   iterative process

                5.   Design phase                     e.   fit the model

                                             S CTI N B (Subjective Type Questions)
            A    Short ans er type  uestions
                1.   What is the purpose of Data Exploration?
              Ans.    After gathering data, processes such as

                     •   data cleaning to locate missing values
                     •   eliminating worthless data (erroneous samples and outliers)
                     •  performing basic statistical analysis such as drawing graphs (or any other visual representation) and comparing
                        different properties of the data set, are carried out. The initial insights gained help to get an understanding of
                        the data and later on, help in algorithm selection, metrics choice, etc. It is useful to see which elements are more
                        essential and what the overall trend of the data is.
                2.   During which stage of the AI model cycle should we take care of AI Ethics? What are the challenges to this process?
              Ans.   During the development of the AI model (modelling phase), care should be taken that the programmer is including
                     data, instructions, etc. that provides protection of all liberties for all citizens as per the fundamental rights of the
                     country. There are two main challenges—One is getting access to high-quality and standardised datasets and second
                     is being able to find skilled programmers who can develop reliable and high-quality machines.


                                                                                         M ode l  L i f e cycl e  163
   160   161   162   163   164   165   166   167   168   169   170