1
Q
Key characteristics of critical thinkers.
A
Analytical, Curious, Research-minded, Open-minded, and Logical
2
Q
What is critical thinking?
A
A process for solving problems and making informed decisions
3
Q
To make data-driven decisions, you follow a process:
A
Understand your goals, ask questions, gather data, study the data to find patterns, make a choice based on the patterns and trends, and evaluate the results.
4
Q
What is data-driven decision-making?
A
Using data to guide your choices and make informed decisions
5
Q
A process for solving problems and making informed decisions
A
Critical thinking
6
Q
Using data to guide your choices and make informed decisions
A
Data-driven decision-making
7
Q
How Does Machine Learning Work?
A
Machine learning is a subset of artificial intelligence that automates data-driven predictions and decisions. It works by studying patterns in the data provided and using that information to detect certain relationships or trends.
8
Q
It works by studying patterns in the data provided and using that information to detect certain relationships or trends.
A
Machine Learning
9
Q
Machine learning utilizes __ to enable computers to learn from the data without requiring instructions on what data should be analyzed
A
Algorithms
10
Q
Which type of question typically seeks further clarification or additional details?
A
Probing questions
11
Q
Why is ethical questioning important in decision-making with AI?
A
12
Q
These graphical representations help you visualize and evaluate decision choices, outcomes, and probabilities
A
Decision trees
13
Q
This decision matrix helps you compare multiple alternatives by quantifying criteria and evaluating their impact.
A
Pugh matrix
14
Q
__ is a decision-making tool you use to evaluate and compare multiple alternatives or options based on a set of criteria
A
Pugh matrix
15
Q
__ are trained with labeled data sets, which allow the models to learn and grow more accurate over time
A
Supervised machine learning models
16
Q
__ a program looks for patterns in unlabeled data
A
unsupervised machine learning
17
Q
__ trains machines through trial and error to take the best action by establishing a reward system. Reinforcement learning can train models to play games, or train autonomous vehicles to drive, by telling the machine when it made the right decisions, which helps it learn over time what actions it should take.
A
Reinforcement machine learning
18
Q
What is the primary function of a Pugh matrix in decision-making?
A
Comparing multiple alternatives
19
Q
What does ensuring fairness in AI models involve?
A
Acknowledging and addressing biases
20
Q
How is AI influencing government policies and decisions?
A
By providing data-driven insights to inform policies
21
Q
How do ethical considerations impact AI-powered healthcare decisions?
A
22
Q
What’s the difference between data-driven decision-making and traditional decision-making?
A
Data-driven decision-making is based on data analysis, while traditional decision-making is based on intuition and personal experience.
23
Q
True or false: AI is a technology that enables machines to learn and perform tasks that would normally require human intelligence.
A
True
24
Q
Data collection
A
is the process of gathering data from various sources.
25
Data labeling
is assigning tags or labels to data to make it more easily searchable and analyzable.
26
Data cleaning
is the process of removing or correcting errors and inconsistencies in the data to improve its quality and accuracy.
27
__ collect data from a group of people using a set of questions. They can be conducted online or in-person, and are often used to collect data on customer preferences and opinions.
Surveys
28
__ collect data from individuals through one-on-one conversations. They can provide more detailed data than surveys, but they can also be time-consuming.
Interviews
29
__ collects data by watching and listening to people or events. This can provide valuable data on customer behavior and product interactions.
Observations
30
__ collects data from websites using software tools. It can be used to collect data on competitors, market trends, and customer reviews.
web scraping
31
Data can be classified into three main categories:
structured, unstructured, and semi-structured.
32
__ occurs when the model is too complex and fits the training data too closely, resulting in poor generalization.
Overfitting
33
__ occurs when the model is too simple and does not capture the underlying patterns in the data.
Underfitting
34
__ occurs when the model is trained on data that is not representative of the real-world population.
Bias
35
While machine learning is a powerful tool for solving a wide range of problems, there are also limitations to its effectiveness including overfitting, underfitting, and bias.
Overfitting, underfitting, and bias
36
__ is a type of machine learning that trains a model to make predictions or decisions based on data.
Predictive AI
37
__ type of machine learning that creates new content, such as images, videos, or text, based on a given input
Generative AI
38
True or false: Ensuring data privacy and confidentiality is only important at the data collection stage.
39
What is GDPR?
The General Data Protection Regulation (GDPR): A set of regulations that apply to all companies that process the personal data of European Union citizens.
40
What is CCPA?
The California Consumer Privacy Act (CCPA): A set of regulations that apply to companies that do business in California and collect the personal data of California residents.
41
What is HIPPA?
The Health Insurance Portability and Accountability Act (HIPAA): A set of regulations that apply to healthcare organizations and govern the use and disclosure of protected health information in the United States.
42
What is EU AI Act?
European Union Artificial Intelligence Act (EU AI Act): Comprehensive AI regulations banning systems with unacceptable risk and giving specific legal requirements for high-risk applications.
43
__ is when a person’s tendency to process information is done by looking for information that is consistent with their existing beliefs while disregarding or ignoring conflicting evidence
Confirmation bias
44
hat is data bias?
Data bias happens when the information you have is not a true reflection of what it should be.
45
Which technique used to identify data bias involves seeking input from colleagues or experts?
46
Using descriptive statistics, such as means, medians, and standard deviations, can reveal potential discrepancies or outliers in the data.
Statistical Analysis
47
Visualizing data through graphs, charts, or histograms can provide valuable insights into data bias. Visualization techniques, like box plots or scatter plots, can help identify patterns or anomalies that can indicate bias.
Data Visualization
48
Exploring the data by conducting thorough data checks and examinations can help identify potential sources of bias.
Data exploration
49
Comparing the dataset with external or independent sources of information can provide an additional means of evaluating bias.
External Validation
50
Seeking input and feedback from colleagues, experts, or peers can enhance the evaluation of data bias. Collaborative efforts can help identify biases that might be overlooked individually and provide diverse perspectives.
Peer Review and Collaboration
51
In which domain can biased AI algorithms impact student performance?
Education
52
How does data bias affect AI applications?
It can lead to discriminatory outcomes.
53
How does data bias affect credit ratings?
It can contribute to financial disparities in our society.
54
How does data bias affect media?
It can amplify existing biases and contribute to information bubbles.
55
How does data bias affect the criminal justice system?
It can have profound implications in the criminal justice system.
56
How does data bias affect healthcare disparities?
It can contribute to disparities in diagnosis and treatment.
57
Bias Detection Algorithms
Employ algorithms to automatically detect and quantify bias in datasets, providing insight into potential sources of that bias.
58
Fairness Metrics
Develop fairness metrics to evaluate the impact of bias on AI models and set thresholds to ensure fairness.
59
Feedback Mechanisms
Collect feedback from end-users and stakeholders to help identify bias and fairness concerns.
60
Responsible Data Collection Practices
Ensure that your data collection methods and sampling techniques are unbiased.
Respect individuals’ privacy rights and obtain informed consent for the collection and use of personal data.
61
Regular Audits and Reviews
Conduct periodic audits and reviews of data collection and preprocessing practices for bias that emerges over time.
62
Monitoring and Improvement
Monitor and improve your AI systems continuously so that you can assess the performance and fairness of your models and make adjustments as needed.
63
Comprehensive Data Analysis
Conduct a thorough analysis of your training data to identify bias that can impact AI decision-making.
Conduct statistical analysis to quantify bias in your data.
Ensure that your training data is diverse and representative of the real-world population. Include data from different demographic groups, geographic locations, and socioeconomic backgrounds.
64
Fairness-Aware Machine Learning
Implement algorithms that explicitly consider fairness as a factor in making predictions.
Example: Use fairness-aware algorithms for loan approvals to ensure equal treatment for all applicants.
65
Fairness Constraints or Regularization
Apply fairness constraints or regularization techniques during model training to reduce bias.
Example: Add constraints to ensure equal false-positive rates across different racial groups in predictive policing models.
66
Fairness Audits and Evaluations
Conduct audits and evaluations to assess the fairness of AI systems and identify potential sources of bias.
Example: Evaluate the fairness of an AI-based credit scoring system through audits and statistical analyses.
67
Resampling
Balance the representation of different groups through oversampling or undersampling.
68
Fairness-Aware Data Splitting
Ensure fairness in training data, validation data, and test data through appropriate splitting.
69
Oversampling/Undersampling
Balance the representation of different groups in the training data.
70
Bias Mitigation Algorithms
Apply algorithms during data preprocessing and during model training to mitigate bias in the training data.
71
Data Augmentation
Increase the quantity and diversity of your training data by augmenting it with other data.
Use external datasets that provide diverse perspectives and representation.
Generate synthetic data to help balance the representation of varied groups.
72
Data Anonymization
Implement proper techniques to protect privacy and reduce bias during data anonymization.
73
What’s the goal of employing bias detection algorithms in AI systems?
To automatically detect and quantify bias in datasets
74
What is responsible AI?
set of practices that ensure AI systems are designed, used, and deployed ethically and legally
75
If the key benefit is Data driven policy development and improved outcomes. The AI application is:
Data Anlaysis for policy insights
76
If the key benefit is real-time public sentiment analysis for policy adjustment. The AI application is:
public opinion analysis
77
If the key benefit is Enhanced decision-making through trend forecasting. The AI application is:
Predictive analytics for decision-making
78
If the key benefit is Streamlined services and cost reduction. The AI application is:
Automation in public services
79
The “black box” nature of AI has been a concern in decision-making processes. AI models should be designed to be __________________. Business users should be able to understand how AI reaches its conclusions, providing a sense of trust and accountability.
Explainable and Transparent
80
The importance of ensuring fairness and upholding ethical guidelines cannot be overstated. _________________ can arise when AI systems are not developed and used responsibly.
Bias, discrimination, and ethical dilemmas
81
Healthcare decision support uses:
AI models assist in disease diagnosis and personalized treatment planning, improving patient care.
The Type of AI model used for this application is:
Supervised
82
AI in Autonomous vehicles use what type of learning
Supervised, Unsupervised, and Reinforcement Learning
83
This type of AI trains machines through trial and error to take the best action by establishing a reward system. Reinforcement learning can train models to play games, or train autonomous vehicles to drive, by telling the machine when it made the right decisions, which helps it learn over time what actions it should take.
Reinforcement machine learning
84
This type of AI program looks for patterns in unlabeled data. It can find patterns or trends that people aren’t explicitly looking for. For example, an unsupervised machine-learning program can look through online sales data and identify different types of clients making purchases.
Unsupervised Machine Learning
85
This type of AI trains machines through trial and error to take the best action by establishing a reward system. It can train models to play games, or train autonomous vehicles to drive, by telling the machine when it made the right decisions, which helps it learn over time what actions it should take
Reinforcement Machine Learning
86
Generative AI uses:
Supervised, unsupervised and reinforcement machine learning
87
This is a decision-making tool you use to evaluate and compare multiple alternatives or options based on a set of criteria. It gives you a systematic approach to make informed decisions by quantifying and ranking different options. The method was developed by a British engineer and design theorist. This method is a valuable tool in fields such as engineering, product design, and business management.
The Pugh Matrix
88
This decision making tool gives you a systematic approach to making choices or decisions based on a set of criteria. It provides a structured way to evaluate different options and their potential outcomes, helping you select the most suitable course of action. It consists of nodes and branches, where nodes represent decisions or choices, and branches represent possible outcomes or consequences.
A decision tree
89
This relies on gut feelings and personal experiences. It is often used when quick decisions are required or when information is incomplete.
Intuitive decision-making
90
This model recognizes that in real-life scenarios, decision-makers often operate with limited information and cognitive constraints. They make decisions that are “good enough” rather than seeking the optimal choice.
Bounded rationality model
91
This model assumes that individuals make logical and consistent decisions by weighing all possible alternatives and selecting the one with the highest utility. Rationality is the cornerstone of this model, and it seeks to make the optimal choice.
Rational decision-making model
92
Why is ethical questioning important in decision-making with AI?
It helps explore and navigate the moral dimensions of choices
93
These questions subtly influence the respondent's perspective, often steering them toward a specific answer or viewpoint.
Leading questions
94
This type of question seek further clarification or additional details, and help you delve
deeper into a topic or issue. This type of question usually comes up after initial
understanding of the topic.
Probing questions
95
Typically answerable with specific numerical answers or a simple “yes” or “no”, “
Blue.”, or “The Great Fire of London”. These questions are useful for gathering specific,
quantifiable data. They are often used in multiple-choice tests or surveys, especially
when computers are used to process the test or survey.
Close ended questions
96
These questions encourage thoughtful and detailed responses, allowing respondents to
express their viewpoints more fully. Use this type of question when you need detailed
explanations behind the concept.
Open ended questions
97
Bill Gates mentions many advantages of AI throughout the video. This is Bill's biggest motivation for pursuing AI. What is it?
Reducing scarcity
98
Bill Gates mentions this category of worker that is most vulnerable to the advances of AI. What is it?
White collar workers
99
How many parameters are there in ChatGPT 4?
How many parameters are there in the human brain?
1 trillion and 100 trillion
100
In her lecture, Dr. Lapata uses a tree as an example of how building parameters improves the capabilities of ChatGPT4. At approximately what level of parameters can ChatGPT begin to be able to provide joke explanations.
(Note: Joke Explanations are visible for a while as the "tree" expands. At what point do you first see it emerge)
220 billion
TIS 7060
flashcards
Decks in class (7)
# Cards
