Case study 2025

What is the main issue with the chatbot implemented by RAKT?

• It is too expensive to maintain.

• It is not user-friendly.

• It cannot understand any customer queries.

• It has a slow response time.

Which of the following is a key area to improve the chatbot's performance?

• Limiting the chatbot's functionalities.

• Increasing the number of employees.

• Improving the chatbot's architecture.

• Reducing the number of customer queries.

What does the term 'latency' refer to in the context of chatbots?

• The amount of data processed.

• The speed of response time.

• The accuracy of responses.

• The complexity of the language model.

Which type of neural network is used by RAKT's chatbot?

• Convolutional Neural Network (CNN)

• Recurrent Neural Network (RNN)

• Generative Adversarial Network (GAN)

• Feedforward Neural Network

What is the purpose of the 'self-attention mechanism' in transformer neural networks?

• To increase the response time.

• To reduce the size of the dataset.

• To simplify the architecture.

• To capture relationships between words.

Which of the following is NOT a type of dataset bias mentioned in the text?

• Cognitive bias

• Confirmation bias

• Historical bias

• Linguistic bias

What is the first step in preparing a machine learning model?

• Training the model

• Pre-processing the input data

• Deployment of the model

• Evaluating the model

What is the main advantage of using Long Short-Term Memory (LSTM) networks?

• They are faster than traditional RNNs.

• They require less data.

• They can handle long-term dependencies better.

• They are easier to train than RNNs.

Which of the following is a challenge related to ethical considerations in chatbot use?

• Data privacy and security

• Improving response time

• Enhancing linguistic features

• Increasing processing power

What is the role of 'natural language understanding' (NLU) in chatbots?

• To generate random responses.

• To analyze customer satisfaction.

• To transform unstructured text into machine-actionable information.

• To limit the number of queries.