Automated Machine Learning for Beginners (Google & Apple)

Discover how automated machine learning automates machine learning workflow, making artificial intelligence accessible to non-data scientists, and how AutoML compares with standard machine learning, its limitations, and the five-step pipeline.

Learn the five-step machine learning pipeline for AutoML, from a clear problem statement and data preparation to model training, deployment, and predictions using Vertex AI and CreateML.

Train a single-label image classification model in Vertex AI by preparing a labeled airplane and car dataset and running AutoML training on a structured train and test set.

Learn to train an image classification model with Vertex AI, deploy to an endpoint, test with images, and undeploy to manage credits, all without coding.

Annotate images with bounding boxes to create object detection data in Vertex AI. Build an AutoML model to recognize dog breeds using labeled bounding boxes, CSV annotations, and cloud storage.

Train and evaluate an object detection model, report average precision around 70 percent, review predictions and misses, deploy to an endpoint, and test with uploaded images.

Explore training an image classification model with CreateML in an AutoML lab, learning how models distinguish animals from non‑animals and cats from dogs, and prepare for mobile deployment.

Learn to train and evaluate a single-label image classification model with CreateML, organizing dataset structure with labeled airplane and automobile images in train and test folders.

Explore object detection by annotating images with bounding boxes using the Oxford Pet Dataset, featuring 37 categories and about 200 images per class.

Use IBM Cloud Annotation to create bounding-box labels for Beagle and Shiba, export annotations to CreateML, and train an object detection model with customizable parameters, a process taking an hour.

Evaluate the object detection model by comparing training, validation, and testing results, noting that test images differing from training data can lower performance, as CreateML classifies pets.

Train a custom sentiment analysis model with Vertex AI AutoML by preparing a text dataset, uploading it to storage, and training, evaluating, and deploying predictions for tweets about U.S. airlines.

Explore evaluating a trained sentiment analysis model, view the evaluation report and confusion matrix, deploy and test with sample texts, then undeploy and conclude the lab.

Evaluate a document classification model by checking precision and misclassifications, deploy to an endpoint, and test with sample CNN articles on politics and space to save time.

Learn to build a text classification model for sentiment analysis with Apple's CreateML, covering data preprocessing, classification parameters, and training workflow using tweets about U.S. airlines.

Learn sentiment analysis by preparing an airline tweet dataset through preprocessing, selecting text and sentiment columns, labeling data for CreateML, and building a CreateML text classification project in Xcode.

Upload training data in CreateML, review class counts, and select transfer learning to train a sentiment analysis model. Evaluate precision and recall, test samples, and export a 1.5 megabyte model.

Train a text classification model with CreateML to assign labels to documents by subject and content, learning dataset structuring, training, and evaluation within a macOS and Xcode environment.

Organize your document classification dataset with train and test folders, where each label is a subfolder of .txt documents, then create a CreateML text classification project in Xcode.

Train a document classification model using transfer learning with 10 categories and 990 texts, evaluate precision and recall, and test predictions before exporting the 1.5 megabytes model.

Train a regression model on tabular data with Vertex AI AutoML tables to predict house sale prices, evaluate with mean absolute error, and deploy for inference.

Learn to build a tabular regression model in CreateML to predict sale price from eight features of the Boston dataset, training with HPP_Train.csv and evaluating with HPP_Test.csv.

Prototype quickly to validate ideas and pitch effectively by building a conceptual, clickable AI alarm clock prototype in Figma, using paper, digital, and native prototyping stages.

Prototype a mobile alarm clock app in Figma by benchmarking Android and iOS designs, then build an AI alarm clock prototype with settings, add alarm, and activation controls.

Deploy a Vertex AI image classification model to an Android app using Android Studio, export a TF Lite model, and integrate it into a sample app with a provided dictionary.txt.

Set up Android Studio, create an empty activity project, and run it; use AVD Manager to create a Pixel 2 virtual device with system image and test with hello world.

demonstrates building an android app with text view, button, and image view in android studio, using setContentView to display the layout and wire the button to change text and image.