Machine Learning Associate ยท MLOps, ML runtimes and AutoML
Databricks AutoML
How AutoML selects models and features, what it produces (best run, notebooks for every trial, data exploration notebook) and the advantages it brings to model development.
Key points
- AutoML prepares your dataset, trains and tunes many models, and picks the best by a metric.
- It generates a notebook per trial, a data exploration notebook and a best trial notebook.
- You can review, reproduce and extend that work.
- Supports classification, regression and forecasting.
Source: https://docs.databricks.com/aws/en/machine-learning/automl/
Practice questions
Which problem types does Databricks AutoML support?
- Classification, regression and forecasting
- Clustering and anomaly detection only
- Reinforcement learning and recommendation
- Image segmentation and speech recognition
Show answer
Answer: A. Classification, regression and forecasting
AutoML covers classification, regression and forecasting. Clustering and reinforcement learning are not supported problem types.
Which algorithms does AutoML try for classification problems?
- Random forests
- XGBoost
- LightGBM
- Prophet
- k-means
Show answer
Answer: A. Random forests, B. XGBoost, C. LightGBM
Classification trials use algorithms such as random forests, XGBoost and LightGBM. Prophet is a forecasting algorithm, and k-means is clustering, which AutoML does not cover.
True or false: AutoML only returns a trained model and hides the code used to produce it.
Show answer
Answer: False
False. AutoML generates a source-code notebook for each trial so you can inspect, reproduce and modify the work.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
