DLOps(Deep Learning Operations) — Learn free

Self-check

Questions 1–10 of 10

1. Which statement summarises the difference between Machine Learning and Deep Learning use cases, as needed to understand DLOps?
Question 1A. Machine Learning is for images; Deep Learning is for tablesB. Machine Learning needs no training data; Deep Learning doesC. Both handle the same use cases but Deep Learning is fasterD. Machine Learning is for simpler, structured and semi-structured use cases; Deep Learning is for complex, unstructured use casesE. Machine Learning is only for text; Deep Learning is only for audio
2. To count the cars, bikes and buses in a set of pictures, which kind of model is needed?
Question 2A. Deep Learning models that can recognise objects in picturesB. Machine Learning models over rows and columnsC. A SQL query on a feature tableD. A data mining model on a data warehouseE. A statistical regression
3. Why can a Machine Learning model not count vehicles when given pictures rather than a table?
Question 3A. Because pictures are too large to loadB. Because Machine Learning models require CI/CDC. Because Machine Learning models normally work on structured or semi-structured datasetsD. Because pictures are not stored in HadoopE. Because Machine Learning cannot count above 100
4. How does the Data Selection step differ between DLOps and MLOps?
Question 4A. DLOps selects smaller datasetsB. DLOps skips Data Selection entirelyC. MLOps selects unstructured data while DLOps selects structured dataD. Data Selection is identical in bothE. DLOps selects unstructured data whereas MLOps selects structured data
5. Which Model Development step is described as not relevant in DLOps because of unstructured datasets?
Question 5A. Problem DefinitionB. Data WranglingC. Model PreparationD. Model ValidationE. Data Selection
6. Which DLOps phase is described as much more difficult in the Deep Learning world?
Question 6A. Problem DefinitionB. Model RegistrationC. Smoke TestD. Data IngestionE. Prediction Serving
7. In the Deep Learning world, what proportion of models is said not to go to production?
Question 7A. 25%B. 50%C. 90%D. 10%E. 70%
8. To which domain do DLOps and MLOps belong?
Question 8A. Data ScienceB. Network EngineeringC. Data WarehousingD. IT Service ManagementE. Database Administration
9. Which step is flagged as one that "might not be relevant" in DLOps?
Question 9A. Model TrainingB. Data ExplorationC. Deployment to StagingD. Acceptance TestingE. Model Evaluation
10. In the vehicle-counting example, which attributes allow a Machine Learning model to identify cars, bikes and buses from a structured dataset?
Question 10A. Colour, brand and priceB. Pixel values of each photographC. Engine size and fuel typeD. Width, length and height of each vehicleE. Registration number and owner