Page 457 - Ai_V3.0_c11_flipbook
P. 457

Answer any 3 out of the 5 questions in 50-80 words each.                                         3x4=12
                   17.  Give brief answers for the following:                                                           4
                       a) If you were designing a robot to sort recyclable items like glass, plastic, and paper, what type of learning
                         would be used to help the robot?

                       b) Any analysis requires the collection and proper format of a vast amount of data. Rohan has stored the
                         information in a delimited file that stores tabular data that is separated by a comma. Which type of file is
                         Rohan using?

                       c)  What do you call the project in which you must research a topic, deeply understand the subject matter, and
                         integrate all the knowledge gathered about a particular subject to develop a solution to a problem?
                       d) Asmita is developing an AI-driven recommendation system for a retail e-commerce platform. What type of
                         machine learning method might she have used to identify groups of similar users or products based on their
                         browsing behaviour?
                   18.  Define Design Thinking. Explain briefly the various steps of Design Thinking.                   4
                   19.  Though AI has gained a lot of popularity, there are still various ethical concerns surrounding it: Explain.   4

                   20.  Consider the given scenario and answer the following questions:                                 4
                       A technology company has been contracted to develop a sophisticated facial recognition system intended for
                       law enforcement applications. This system is designed to assist in identifying suspects by matching their facial
                       features against a database. Initially recognised for its high level of accuracy and efficiency in processing and
                       identifying faces, the system, however, has come under scrutiny after reports emerged that it disproportionately
                       misidentifies people of colour. Investigations revealed that the misidentification rate for individuals of colour was
                       significantly higher than for white individuals, leading to a higher incidence of wrongful detentions and arrests
                       among these communities. Further analysis indicated that the training data used to develop the algorithm lacked
                       sufficient diversity, predominantly featuring the faces of white individuals. This oversight has not only raised
                       questions about the racial bias inherent in AI technologies but also about the ethical implications of deploying
                       such systems in sensitive areas like law enforcement without thorough vetting for fairness and impartiality.

                       a) What ethical problems are evident in this scenario?
                       b) What is one ethical approach to mitigate the bias found in the facial recognition system without
                         compromising its accuracy?
                       c)  How does the lack of diversity in training data contribute to the problem of algorithmic bias in this scenario?
                       d) What measures can be taken to ensure that the deployment of AI systems in law enforcement is both ethical
                         and effective?

                   21.  How does classification model work?                                                             4

























                                                                                          Model Test Paper-1    455
   452   453   454   455   456   457   458   459   460   461   462