“Machine intelligence is the last invention that humanity will ever need to make.”
Nick Bostrom
Automated Machine Learning (AutoML) is the process of automating the end-to-end process of applying Machine Learning to real-world problems. In a typical ML application, experts must apply the appropriate data pre-processing methods, feature engineering, feature extraction, and feature selection to make the data set most accessible for ML. Following these pre-processing steps, practitioners must then perform the algorithm selection and hyper-parameter optimization to maximize the predictive performance of the final ML model. Since many of these steps often go beyond the capabilities of laypersons, AutoML has been developed as an artificial intelligence-based solution to the ever-growing challenge of applying ML. Automating the end-to-end process of applying ML offers the benefits of producing more straightforward solutions, faster creation of these solutions, and models that often outperform hand-designed models.
Oracle AutoML UI
AutoML User Interface (AutoML UI) is an Oracle Machine Learning interface that provides you no-code automated machine learning modelling. When you create and run an experiment in AutoML UI, it performs automated algorithm selection, feature selection, and model tuning, thereby enhancing productivity as well as potentially increasing model accuracy and performance.
The following steps comprise a machine learning modelling workflow and are automated by the AutoML user interface:
- Algorithm Selection: Ranks algorithms likely to produce a more accurate model based on the dataset and its characteristics, and some predictive features of the dataset for each algorithm.
- Adaptive Sampling: Finds an appropriate data sample. The goal of this stage is to speed up Feature Selection and Model Tuning stages without degrading the model quality.
- Feature Selection: Selects a subset of features that are most predictive of the target. The goal of this stage is to reduce the number of features used in the later pipeline stages, especially during the model tuning stage to speed up the pipeline without degrading predictive accuracy.
- Model Tuning: Aims at increasing individual algorithm model quality based on the selected metric for each of the shortlisted algorithms.
- Feature Prediction Impact: This is the final stage in the AutoML UI pipeline. Here, the impact of each input column on the predictions of the final tuned model is computed. The computed prediction impact provides insights into the behaviour of the tuned AutoML model.
