We all know that Machine learning (ML) is the process that enables computers to learn without explicit programming based on historical data. This can lead to intelligent actions that have not be explicitly coded in the system thereby termed Artificial Intelligence (AI)
Let’s start with the basics and understand the key steps involved in the process
- Dataset – Predefined data examples that will be used to create the model
- Labels – Categories defined for the each data item in the data set
- Training – Process of teaching the model using data set about the categories so that it can recognize when there is a input data
- Models – A classification system that is trained using the data set so that it can predict when there is input data
- Prediction – Output provided by a trained model that says how closely the input matches data in the dataset
Einstein Platform Services is a set of APIs made available for developers to perform these steps and has two broad categories – Vision and Language
Einstein Vision is further categorized as follows
- Einstein Image Classification – Identify what a particular image is
- Einstein Object Detection – Recognize and count multiple distinct objects within an image
There are prebuilt models available as well as APIs available to build the custom ones
Following are prebuilt Vision models
- Food Image Model – e.g. beer, broccoli
- General Image Model – e.g. Zebra, apron
- Scene Image Model – e.g. bank interior, stadium
- Multi-Label Image Model – e.g. camera, starfish
Einstein Language is further categorized asÂ
- Einstein Sentiment – Predict a given text implies as positive, negative or neutral
- Einstein Intent – Recognize what user is try to ask for
Prebuilt model for language is the Community Sentiment Model that you can use as long as you have a valid JWT token. This model has three classes:
• positive• negative• neutral
Trying out the APIs with Postman client is the best way to test the waters, be sure to checkout the Trailhead as well