Definitions

The definition of key concepts

Item

description

Annotation

The process of labelling the data available in various formats like text, video or images.

Firm / Business

A concept object which can manage one or more projects.

Business Model

A concept model that pre-defines fields from which the prediction will extract values.

MultiClass-Classification

A classification task with two or more classes with the assumption that each sample is assigned to ONLY ONE label - e.g. A document can be either an insurance receipt or water bill, but not both at the same time.

MultiLabel-Classification

A classification task with two or more classes with the assumption that each sample is assigned to MULTIPLE labels - e.g. A document can cover topics of both religion and politics at the same time.

Free form (Data Extraction)

One of project types in which user can freely define interested fields for AI to learn and extract values from predicated documentation.

Project Type

The type of AI task the created project aims to tackle:

  • MultiClass-Classification
  • MultiLabel-Classification
  • Free Form (Data Extraction)
  • Various pre-built models

Pre-built Model

A model that is pre-trained based on selected data and annotation. User can start predicting documentation directly by saving the LABEL(annotation) and TRAIN processes whenever the pre-built model fits the business need.

AI-assistant / AI chat bot

Easily create a custom AI assistant or chatbot on our platform, powered by OpenAI (with future options to choose between OpenAI and Claude). Simply type your instructions to tailor your AI assistant to your needs. For advanced functionality, you can integrate with your own APIs through our collaborative support. Additionally, link your AI assistant to a project within AI Studio, allowing it to leverage stored documents as its knowledge base. Integration is seamless—share the AI assistant as a web link or embed it directly into your website with a simple UI component.