ISMN Chp 14
Weak AI is already powerful enough to make a dramatic difference in human life.
False
E-mail spam detection is an example of _____ classification.
binary
Weak AI is hypothetical.
false; strong
Experts have a low concern for burglar bots.
true
_______ learning is a type of machine learning that searches for previously undetected patterns in a data set with no pre-existing labels and with minimal human supervision.
unsupervised
_ is a startup that develops machine learning-powered, voice-activated platforms with technology that is __.
Apprente; sound-to-meaning
BlackSwan is an AI company whose product, Element, integrates _____ to address data acquisition, insight discovery, and predictions.
Big Data and machine learning
Conversion prediction is an example of _____ classification.
Binary
_______ is an example of semi-supervised learning.
Classifying text documents
__ require human experts; __ must be formally structured in the form of rules.
Expert systems; expert systems
A hyperparameter in a neural network consists of software that has one or more weighted input connections, a bias function, and activation function, and one or more output connections.
False
A virtual twin is a virtual, digital model of a machine or a person created from real-time and historical data.
False
Chatbots do not require a knowledge base; they can offer basic problem solving.
False
Convolutional neural networks can be used to create deepfakes ranging from image manipulation to news embellishments.
False
Customer churn has so many factors associated with it that machine learning has not proven to be particularly effective yet.
False
Expert systems always replace human decision makers.
False
Expert systems are an application of AI that provide systems with the ability to automatically learn and improve from experience without being explicitly programmed.
False
Frito-Lay paints chips and then uses computer vision to determine the texture of the chip.
False
Information agents and predictive agents perform the same sorts of tasks.
False
Linear regression refers to a predictive modeling problem in which the system generates a class label for a given set of input data.
False
Neural networks rarely make mistakes given the amount of data used to train them.
False
Supervised machine learning is an unstructured combination of data and a computer algorithm that produces answers.
False
The best time to use unsupervised learning is when an organization does not have data on desired outcomes.
False
_____ is a result that indicates that a given condition exists when it in fact does not.
False positive
The machine learning-powered chatbot by Amazon answers more than 350 million customer inquiries per day, successfully understanding more than 95 percent of them.
False; Alibaba
Chick-Fil-A implemented Dynamic Yields personalization functionality on electronic ordering boards at its drive-through locations to deliver personalized offerings, to appeal more to customers, and to help the kitchen during periods of peak activity.
False; Mcdonalds
Natural image processing refers to the ability of information systems to identify objects, scenes, and activities in images.
False; computer vision
Experts are moderately concerned with burglar bots.
False; low concern
Classification and SEM analysis are important techniques for supervised learning.
False; not SEM
Sephora Virtual Artist is the augmented reality tool that allows customers to try on thousands of makeup products.
False; virtual reality
Which of the following is NOT reinforcement learning?
Labeling spam
Edge technology from Kroger and Just Walk Out technology from Amazon leverage ____-driven technologies.
Machine learning
_______ is a structured combination of data and a computer algorithm that produces answers.
Not nontraditional prog
_____ is a capability performed relatively better by AI than human intelligence.
Total cost of knowledge
A digital twin is a virtual, digital model of a machine or a person created from real-time and historical data.
True
A node in a neural network consists of software that has one or more weighted input connections, a bias function, and activation function, and one or more output connections.
True
An smart factory is a highly digitized operation that continuously collects and shares data through connected machines, devices, and production systems.
True
Any time a learning system uses neural networks, it is called a deep learning system.
True
Chatbots are not considered intelligent.
True
Chatbots are useful in call centers.
True
Current AI systems are considered weak AI.
True
Deep learning systems do not need to be exposed to labeled historical or training data.
True
DevOps bots are still under development; they are machine learning-powered bots that assist in all stages of the software development process.
True
Drones share jobs with humans on the factory floor.
True
Even if human experts can explain their entire reasoning process, it may still be impossible to automate that process.
True
Expert systems are computer systems that attempt to mimic human experts by applying expertise in a specific domain.
True
Expert systems require human experts.
True
Generative adversarial networks can be used to create deepfakes ranging from image manipulation to news embellishments.
True
Generative adversarial networks perform well in applications where filling in missing or incomplete data may be required.
True
If spam e-mails are labeled as safe, this is called data poisoning.
True
Intelligent agents do not simply provide answers from a knowledge base; they can offer basic problem solving.
True
Machine learning algorithms learn from ingesting vast amounts of data and by adjusting hyperparameters and parameters.
True
Machine learning is sometimes better at recognizing patterns and relationships than people.
True
Machine learning systems used for criminal risk assessment have been found to be biased against people of color.
True
McDonalds implemented Dynamic Yields personalization functionality on electronic ordering boards at its drive-through locations to deliver personalized offerings, to appeal more to customers, and to help the kitchen during periods of peak activity.
True
Speech recognition focuses on automatically and accurately transcribing human speech.
True
The machine learning-powered chatbot from Alibaba answers more than 350 million customer inquiries per day, successfully understanding more than 90 percent of them.
True
There are four types of classification for supervised learning: binary, multi-class, multi-label, and imbalanced.
True
___ is a capability performed relatively better by human intelligence than AI.
Use of sensory experiences
______ has installed cameras at each checkout lane to detect the mood of a customer using recognition technology.
Walmart?
Machine learning systems trained on datasets collected from biased samples can exhibit these biases when they are used; this is called ______.
algorithmic bias
_____ refers to a situation in which a version of the final software package is always ready to be released but is not sent to production before the decision is made to release it.
continuous delivery
_____ is a subset of machine learning in which artificial neural networks learn from large amounts of data.
deep learning
Smart factories do NOT use _____.
expert systems
Which of the following is NOT a type of reinforcement learning?
finding customer segments
Medical diagnostic tests are an example of _____ classification.
imbalanced
Generative adversarial networks are used in which of the type of application below?
improving deep space photography
Generative adversarial networks are used in applications such as _____.
inpainting
_______ is one of the primary techniques in supervised learning.
linear regression
_______ is an application of AI that provides systems with the ability to automatically learn and improve from experience without being explicitly programmed.
machine learning
Deep learning is a subset of __ in which artificial ___ learn from large amounts of data.
machine learning; neural networks
Labeling spam emails as safe is called data ___ and experts consider it a __ concern AI-enabled crime.
poisoning; moderate
A(n) ___ neural network is designed to access previous data such as sequential data or time series data during iterations of input.
recurrent
A(n) ___ neural network is used in applications such as moving a robotic arm, reading a sentence, predicting time series, and composing music.
recurrent
_______ learning is a type of machine learning in which the system learns to achieve a goal in an uncertain, potentially complex environment.
reinforcement