The hidden layer that contains the compressed representation of the input data in an autoencoder is called the ________ layer.
- Bottleneck
- Compression
- Encoding
- Latent
The hidden layer that holds the compressed representation in an autoencoder is the 'Latent' layer, capturing essential features of the input data.
What role do the hidden states in RNNs play in terms of sequential data processing?
- Storing Information Over Time
- Managing Data Loss
- Encoding Input Features
- Updating Weights for Classification
The hidden states in RNNs play a crucial role in storing information over time. They retain memory of past inputs and contribute to the model's ability to process sequential data, making them suitable for tasks with dependencies over time.
Which of the following describes the situation when a model performs well on the training data but poorly on unseen data?
- Bias
- High Variance
- Overfitting
- Underfitting
This situation is known as overfitting, where a model learns to fit the training data too closely but fails to generalize to new, unseen data, resulting in a high error rate.
One of the challenges in DQN is that small updates to Q values can lead to significant changes in the policy, making the learning process highly ________.
- Sensitive
- Efficient
- Predictable
- Robust
The term 'sensitive' in this context refers to the fact that small changes in Q values can have a disproportionate impact on the policy, making it unstable and hard to control.
The multi-armed bandit problem can be viewed as a simplified version of the reinforcement learning problem where the number of ________ is just one.
- Episodes
- States
- Actions
- Rewards
The multi-armed bandit problem simplifies reinforcement learning to just one action, where you need to decide which arm of a bandit to pull.
In a video game with multiple levels and complex interactions, what approach is suitable for training an AI agent optimally?
- Transfer Learning
- Curriculum Learning
- Random Search
- Supervised Learning
Curriculum Learning is ideal for training an AI agent to handle various levels with different challenges. It starts with easy levels, gradually increasing difficulty based on the agent's performance in earlier stages, ensuring effective learning.
The ability of an individual or a group to understand and trust the model's decisions is often tied to the model's ________.
- Explainability
- Complexity
- Accuracy
- Processing speed
Model explainability is essential for understanding and trusting a model's decisions, especially in critical applications like healthcare or finance, where transparency is key for decision-making and accountability.
In the k-NN algorithm, when two classes have a similar number of instances close to a test data point, the choice of an odd 'k' can help to avoid ________
- Ambiguity
- Bias
- Overfitting
- Underfitting
When two classes have a similar number of instances nearby, using an odd 'k' value can help avoid ambiguity in classifying the test data point, as it prevents ties. An even 'k' can lead to ties, making classification less clear.
What term refers to the ability to understand and interpret machine learning model decisions?
- Explainability
- Predictability
- Efficiency
- Generalization
Explainability is the term that refers to the ability to understand and interpret machine learning model decisions. It's crucial for trust and accountability in AI systems.
Which variant of RNN is designed to better capture long-term dependencies in sequence data?
- Vanilla RNN
- Bidirectional RNN
- Long Short-Term Memory (LSTM)
- Gated Recurrent Unit (GRU)
Long Short-Term Memory (LSTM) is designed to capture long-term dependencies in sequences by maintaining cell states that are less susceptible to the vanishing gradient problem.