Consider a self-driving car learning from trial and error in a simulated environment. This is an example of which type of learning?
- Deep Learning
- Reinforcement Learning
- Supervised Learning
- Unsupervised Learning
This scenario exemplifies Reinforcement Learning. In Reinforcement Learning, an agent learns to take actions in an environment to maximize a reward signal. The self-driving car explores different actions (e.g., steering, accelerating, braking) and learns from the consequences in a simulated environment to improve its driving skills.
________ is the problem when a model learns the training data too well, including its noise and outliers.
- Bias
- Overfitting
- Underfitting
- Variance
Overfitting is the problem where a model becomes too specialized in the training data and captures its noise and outliers. This can lead to poor performance on unseen data.
When dealing with a small dataset and wanting to leverage the knowledge from a model trained on a larger dataset, which approach would be most suitable?
- Fine-tuning
- Transfer Learning
- Random Initialization
- Gradient Descent Optimization
The most suitable approach for leveraging knowledge from a model trained on a larger dataset with a small dataset is "Transfer Learning." It involves adapting the pre-trained model to the new task.
In hierarchical clustering, as the name suggests, the data is grouped into a hierarchy of clusters. What visualization is commonly used to represent this hierarchy?
- Bar Chart
- Dendrogram
- Heatmap
- Scatter Plot
A dendrogram is commonly used in hierarchical clustering to visualize the hierarchical structure of clusters, showing the merging and splitting of clusters.
Q-learning is a type of ________ learning algorithm that aims to find the best action to take given a current state.
- Reinforcement
- Supervised
- Unsupervised
- Semi-supervised
Q-learning is a type of reinforcement learning that focuses on finding the best action to take in a given state to maximize cumulative rewards.
When aiming to reduce both bias and variance, one might use techniques like ________ to regularize a model.
- Cross-Validation
- Data Augmentation
- Dropout
- L1 Regularization
L1 regularization is a technique used to reduce both bias and variance in a machine learning model. It does so by adding a penalty term to the model's loss function, which encourages the model to use fewer features, thus reducing complexity and variance. Dropout, Cross-Validation, and Data Augmentation are techniques but are not primarily used for regularization.
What does the "G" in GRU stand for when referring to a type of RNN?
- Gated
- Global
- Gradient
- Graph
The "G" in GRU stands for "Gated." GRU is a type of RNN that uses gating mechanisms to control information flow, making it capable of handling sequences efficiently.
One of the challenges in training deep RNNs is the ________ gradient problem, which affects the network's ability to learn long-range dependencies.
- Vanishing
- Exploding
- Overfitting
- Regularization
The vanishing gradient problem refers to the issue where gradients in deep RNNs become too small during training, making it challenging to capture long-range dependencies.
In the context of the bias-variance trade-off, which one is typically associated with complex models with many parameters?
- Balanced Bias-Variance
- High Bias
- High Variance
- Neither
High Variance is typically associated with complex models with many parameters. Complex models are more flexible and tend to fit the training data closely, resulting in high variance, which can lead to overfitting.
In time series forecasting, the goal is to predict future ________ based on past observations.
- Events
- Trends
- Weather
- Stock Prices
Time series forecasting aims to predict future trends or patterns based on historical data, which can be applied in various fields like finance or weather.