Multiple Choice Flashcards
(43 cards)
A researcher proposes using ReLU activation in the final layer of a binary classification model. What will likely happen?
- The output will not be properly scaled to probability values between 0 and 1
- The model will perform better since ReLU allows unrestricted values
- ReLU will output negative values, making it a better fit for binary classification
- The model will generalize better since ReLU prevents saturation
- The output will not be properly scaled to probability values between 0 and 1
How can schools utilize data literacy?
- By analyzing student performance and improving education strategies
- By avoiding technology in the learning process
- By restricting access to educational data
- By discouraging students from using data
- By analyzing student performance and improving education strategies
What is a key characteristic of an Intelligent System?
- Follows fixed rules
- Cannot adapt to new situations
- Learns from experience and adapts
- Only works with human intervention
- Learns from experience and adapts
A CNN trained on synthetic images fails when tested on real-world photographs. What is the best approach to fix this?
- Fine-tune the model with a dataset containing real-world images
- Increase the depth of the CNN to extract more detailed features
- Reduce the number of convolutional layers to prevent overfitting
- Use grayscale images to reduce dependency on color differences
- Fine-tune the model with a dataset containing real-world images
Why is ReLU preferred over sigmoid for deep neural networks?
- ReLU avoids vanishing gradients by not saturating for positive inputs
- ReLU is always differentiable everywhere
- ReLU uses a probabilistic approach to activation
- Sigmoid outperforms ReLU in all deep learning applications
- ReLU avoids vanishing gradients by not saturating for positive inputs
Why is communication important in data literacy adoption?
- It helps people understand why data literacy matters
- It creates unnecessary discussions
- It complicates the implementation process
- It eliminates the need for training
- It helps people understand why data literacy matters
An organization wants to ensure data fairness. What should they prioritize?
- Ensuring diverse and unbiased data sources
- Using data only from a single demographic
- Avoiding data collection altogether
- Assuming all data is neutral by default
- Ensuring diverse and unbiased data sources
What distinguishes AI from traditional software systems?
- It operates only with pre-defined rules
- It does not require any training
- It adapts and learns over time
- It is strictly hardware-based
- It adapts and learns over time
A school wants to track student performance beyond test scores. What additional data sources should be analyzed?
- Only grades from final exams
- Test scores from four years ago
- Random student opinions without structured feedback
- Attendance, participation, assignment completion, and teacher feedback
- Attendance, participation, assignment completion, and teacher feedback
A company collects large amounts of customer data but struggles to make informed decisions. What is the most likely reason?
- The data is too large to analyze
- The company lacks a data literacy culture and structured interpretation methods
- The company only needs to store data, not analyze it
- Data literacy is only useful for technical teams
- The company lacks a data literacy culture and structured interpretation methods
What are the key modules of data literacy?
- Data Storage, Computation, and Security
- Mathematics, Programming, and Logic
- Interpretation, Clarity, and Storytelling
- Data Mining, Cleaning, and Visualization
- Interpretation, Clarity, and Storytelling
A city government uses data to improve traffic management. What is the best way to ensure accurate decision-making?
- Rely on historical traffic data only
- Ignore public transportation data
- Assume that traffic patterns remain constant year-round
- Collect and analyze real-time traffic patterns, weather conditions, and commuting trends
- Collect and analyze real-time traffic patterns, weather conditions, and commuting trends
How does Leaky ReLU address the shortcomings of ReLU?
- By allowing a small gradient for negative inputs, preventing neuron death
- By improving computational efficiency over ReLU
- By ensuring that all neurons are always active
- By replacing the max function with an exponential function
- By allowing a small gradient for negative inputs, preventing neuron death
Who are “Data Aristocrats” in data literacy personas?
- Beginners in the world of data
- People who do not believe in data literacy
- Individuals who work to establish the right vision and buy-in
- People who only use data for entertainment purposes
- Individuals who work to establish the right vision and buy-in
A CNN model trained for medical image classification performs well on the training set but fails when tested on real-world hospital data. What is the most probable cause?
- The model architecture is too simple
- The training data lacks diversity, leading to poor generalization
- The optimizer used during training is incorrect
- CNNs are not suitable for medical imaging applications
- The training data lacks diversity, leading to poor generalization
What is the first step in implementing a data literacy program?
- Measuring program success
- Cultural learning
- Planning and vision
- Workforce assessment
- Planning and vision
How can data literacy be adopted?
- By training employees and cultivating skills
- By keeping data restricted to management only
- By limiting access to data tools
- By relying only on intuition for decision-making
- By training employees and cultivating skills
What is the role of actuators in an intelligent system?
- Collecting data
- Analyzing user feedback
- Storing knowledge
- Executing system responses
- Executing system responses
What is a common use case of reinforcement learning?
- Image classification
- Data encryption
- Traditional database queries
- Robotics and game playing
- Robotics and game playing
Why is data literacy important?
- It eliminates the need for human analysis
- It makes people better at writing reports
- It only benefits data scientists
- It helps improve decision-making skills
- It helps improve decision-making skills
A government agency wants to ensure transparency in its data reports. What practice should it adopt?
- Only share positive data insights
- Provide access to complete datasets, methodologies, and potential biases
- Publish reports without explanation
- Avoid external review of the data
- Provide access to complete datasets, methodologies, and potential biases
A school wants to adopt a data-driven learning approach. What strategy should they use?
- Collect and analyze data on student engagement, performance, and learning styles
- Ignore data and use the same teaching method for all students
- Focus only on final exam results
- Assume all students learn the same way
- Collect and analyze data on student engagement, performance, and learning styles
What is the definition of data literacy?
- The ability to read books effectively
- The ability to memorize large amounts of information
- The ability to write code in any programming language
- The ability to read, work with, analyze, and argue with data
- The ability to read, work with, analyze, and argue with data
A data analyst detects a significant data outlier in a financial report. What should be their next step?
- Ignore it as an error and continue with the analysis
- Investigate whether the outlier is a legitimate insight or a mistake in data collection
- Remove all outliers to simplify data interpretation
- Base all future financial decisions on the outlier
- Investigate whether the outlier is a legitimate insight or a mistake in data collection
