MIS 375 Exam 2 Questions and
Answers20
Audio Database - ANSWERS -trained with every spoken language, regional accents, relation of
words - not just hearing it but also noting the differences + employment of predictive/
suggestions
used for tasks like speech recognition and audio analysis
Training involves supervised learning with labeled audio data
Ex - hey google! hey alexa! hey siri!, 1-800 phone lines use of AI
Textual Database - ANSWERS -has every language in in, grammar, and translations/relationships
between languages and between words
trained using unsupervised pre-training and fine-tuning on specific tasks
(emphasis on the fact that it takes many years to fully develop such a database, the companies
that have textual neural databases are proprietary - facebook, apple, google)
Ex - Chatgpt & Grammarly
Visual Database - ANSWERS -used for tasks like image recognition and computer vision
,trained by collecting data - for instance google analyzes everything that comes through it, the
login site when you click for objects also help train visual databases (supervised learning with
labeled audio data), potential for predictive picture expansion with AI
Ex - Google, Meta
Why would a private data source be important as opposed to a public data source as a
foundation for a generative AI large language model? - ANSWERS -That main reason to opt for a
private data source instead of a public one for a LLM would be that organizations could create AI
models that are highly customized to their unique needs. So they could train an AI model based
on content that is more relevant and specific to their goals/ requirements and the organization's
specific domain and terminology. They would also have more control over the data and the
model in order to enable greater security measures.
What are the main forms of AI / ML? What are the technologies that are used for each? What
purpose do they serve? - ANSWERS -Decision Tree (Simple Decisions)
Decision Tree (Random Forest/Machine Learning)•
LLM/Generative AI (Knowledge Sifting - Public/Private, Text Generation, Language Translation,
Code Generation, Audio Interaction, Image Interpretation), can use Natural Language Processing
Algorithmic AI (Problem Solving, Autonomous Decision Making, Intelligent Monitoring,
Probability Based Actions, Collaborative Filtering)
EX: Bayesian Networks, Time Series Analysis
What are the three types of neural databases? - ANSWERS -Textual, Visual, Audio
What is prompting? - ANSWERS -providing a specific input or instruction to an AI model to
generate a desired output or response
, the prompt is a text-based or verbal query that a human user provides to the AI model, and it
serves as the starting point for the model to generate a response or complete a task
Common Prompts Used for Prompting - ANSWERS -Simple questions (what is... explain....)
Lists (generate a list for....)
Long-form written pieces (write a tweet/essay/paragraph)
Complex problems (show mathematical equations that lead to)
Feedback (critique this paragraph/essay)
Modify (translate/paraphrase/lengthen this.....)
Instructional guides (how do I..../ dinner ideas with these ingredients.....)
Write and debug code
Conversation and advice
What are the four main divisions of Disney? How could AI be used for each? - ANSWERS -1.
Media Networks (ABC, Disney Channel, ESPN)
AI Use - More important for ESPN but video analysis and tagging to create a more engaging live
sports viewing experience and enhance replays
Answers20
Audio Database - ANSWERS -trained with every spoken language, regional accents, relation of
words - not just hearing it but also noting the differences + employment of predictive/
suggestions
used for tasks like speech recognition and audio analysis
Training involves supervised learning with labeled audio data
Ex - hey google! hey alexa! hey siri!, 1-800 phone lines use of AI
Textual Database - ANSWERS -has every language in in, grammar, and translations/relationships
between languages and between words
trained using unsupervised pre-training and fine-tuning on specific tasks
(emphasis on the fact that it takes many years to fully develop such a database, the companies
that have textual neural databases are proprietary - facebook, apple, google)
Ex - Chatgpt & Grammarly
Visual Database - ANSWERS -used for tasks like image recognition and computer vision
,trained by collecting data - for instance google analyzes everything that comes through it, the
login site when you click for objects also help train visual databases (supervised learning with
labeled audio data), potential for predictive picture expansion with AI
Ex - Google, Meta
Why would a private data source be important as opposed to a public data source as a
foundation for a generative AI large language model? - ANSWERS -That main reason to opt for a
private data source instead of a public one for a LLM would be that organizations could create AI
models that are highly customized to their unique needs. So they could train an AI model based
on content that is more relevant and specific to their goals/ requirements and the organization's
specific domain and terminology. They would also have more control over the data and the
model in order to enable greater security measures.
What are the main forms of AI / ML? What are the technologies that are used for each? What
purpose do they serve? - ANSWERS -Decision Tree (Simple Decisions)
Decision Tree (Random Forest/Machine Learning)•
LLM/Generative AI (Knowledge Sifting - Public/Private, Text Generation, Language Translation,
Code Generation, Audio Interaction, Image Interpretation), can use Natural Language Processing
Algorithmic AI (Problem Solving, Autonomous Decision Making, Intelligent Monitoring,
Probability Based Actions, Collaborative Filtering)
EX: Bayesian Networks, Time Series Analysis
What are the three types of neural databases? - ANSWERS -Textual, Visual, Audio
What is prompting? - ANSWERS -providing a specific input or instruction to an AI model to
generate a desired output or response
, the prompt is a text-based or verbal query that a human user provides to the AI model, and it
serves as the starting point for the model to generate a response or complete a task
Common Prompts Used for Prompting - ANSWERS -Simple questions (what is... explain....)
Lists (generate a list for....)
Long-form written pieces (write a tweet/essay/paragraph)
Complex problems (show mathematical equations that lead to)
Feedback (critique this paragraph/essay)
Modify (translate/paraphrase/lengthen this.....)
Instructional guides (how do I..../ dinner ideas with these ingredients.....)
Write and debug code
Conversation and advice
What are the four main divisions of Disney? How could AI be used for each? - ANSWERS -1.
Media Networks (ABC, Disney Channel, ESPN)
AI Use - More important for ESPN but video analysis and tagging to create a more engaging live
sports viewing experience and enhance replays
