ML System Deployment Quiz

What is the primary focus of ML in production compared to ML in research?

• Improving state-of-the-art models

• Enhancing model interpretability

• Optimizing, automating, or augmenting a business scenario

• Increasing training speed

In the context of ML, what does 'inference' refer to?

• Training the model

• Predicting or making a decision using an ML model

• Preprocessing data

• Evaluating model performance

Which of the following is a deployment target for ML models?

• Virtual machines

• Containers

• Serverless computing

• All of the above

What is one advantage of deploying ML models on virtual machines?

• No need for infrastructure management

• More resource efficiency

• Customizable runtime environment

• Easy to interpret models

Which tool is commonly used to build and deploy containers?

• Docker

• Flask

• Apache Kafka

• Spring Boot

Which deployment method allows for scaling to zero to save computation power?

• Virtual machines

• Containers

• Serverless computing

• Model streaming

What is a disadvantage of using serverless computing for ML model deployment?

• Requires expertise to manage

• Limited to synchronous operations

• Potential resource limitations like RAM or disk space

• High latency for model inference

What does model streaming help handle effectively in ML model deployment?

• Static datasets

• Batch processing

• Streaming data

• Server management

Which of the following is a message broker used in model streaming?

• Docker

• Apache Kafka

• Spring Boot

• Flask

Which deployment target is best for scenarios requiring multiple models to be inferred in an ML system?

• Virtual machines

• Containers

• Serverless computing

• Model streaming