Transformers QUIZ (MCQ QUESTIONS AND ANSWERS)

Convolutional Neural Network (CNN).

Recurrent Neural Network (RNN).

Autoencoder.

Feedforward Neural Network (FNN).

A signal indicating the start of an episode.

A signal indicating the end of an episode.

A signal indicating the quality of an agent's action in an environment.

A signal indicating the learning rate.

Reduced computational complexity.

Improved interpretability.

The ability to learn complex hierarchical features.

Faster training times.

When a model performs well on the training data but poorly on unseen data.

When a model has too few parameters.

When a model performs equally well on both training and validation data.

When a model underfits the training data.

Convolutional Layer.

Recurrent Layer.

Dense Layer.

Pooling Layer.

To increase the size of the training dataset.

To reduce the learning rate.

To increase the number of hidden layers.

To preprocess input data.

ReLU.

Sigmoid.

Tanh.

Softmax.

Gradient Descent.

K-Means.

Principal Component Analysis (PCA).

Decision Trees.

To increase model complexity.

To compute predictions.

To evaluate model performance.

To quantify the error between predicted values and ground truth.

To increase the batch size during training.

To normalize input data.

To reduce computational complexity.

To stabilize and accelerate training by normalizing layer activations.

A type of activation function.

A type of layer.

One complete pass through the entire training dataset.

A type of regularization technique.

The process of transferring model weights from one neural network to another.

The process of transferring knowledge learned from one task to improve performance on another task.

The process of transferring data between neural networks.

The process of fine-tuning a pre-trained model.

Convolutional Neural Network (CNN).

Recurrent Neural Network (RNN).

Feedforward Neural Network (FNN).

Autoencoder.

To introduce non-linearity into the model.

To reduce the dimensionality of the input data.

To compute class probabilities for multi-class classification.

To increase the number of layers in the network.

To convert text into binary code.

To identify keywords in a text.

To split text into smaller units, such as words or subwords.

To translate text into multiple languages.

Image classification.

Sequential data analysis.

Speech recognition.

Reinforcement learning.

To increase the number of neurons in each layer.

To reduce the dimensionality of the input data.

To prevent overfitting by randomly deactivating neurons during training.

To speed up training by skipping some training examples.

A problem where gradients become too large during training.

A problem where gradients become too small during training, hindering learning.

A problem where the learning rate is too high.

A problem where the model converges too quickly.

To reduce the dimensionality of data.

To preprocess input data.

To introduce non-linearity into the model.

To increase the number of layers in the network.

Small number of parameters.

Lack of self-attention.

Computational expense and memory requirements.

Inability to handle sequential data.

Developing low-end devices.

Large dataset analysis.

Computer vision.

Reinforcement learning.

GPT-3.

BERT.

BART.

RoBERTa.

They are not effective for text processing.

They have a small number of parameters.

They are computationally expensive and require large datasets.

They are not suitable for natural language processing.

Using a very small batch size.

Ignoring positional encoding.

Avoiding dropout layers.

Employing early stopping.

Using dropout layers.

Reducing the learning rate.

Increasing the batch size.

Removing the positional encoding.

Image classification.

Reinforcement learning.

Natural language processing tasks.

Numerical simulations.

Self-Attention.

Multi-Head Attention.

Positional Encoding.

Layer Normalization.

Increasing the number of layers.

Reducing computational complexity.

Focusing on different aspects of the input sequence.

Eliminating positional encoding.

Stacked LSTM.

Bidirectional LSTM.

LSTM with attention mechanism.

Unidirectional LSTM.

Paying attention to oneself.

Focusing on the most important words in a sentence.

Learning context in sequential data.

Avoiding overfitting.