Beginner's Guide to Commonly Used Terms in AI & Intelligent Automation

Share on facebook
Share on google
Share on twitter
Share on linkedin

You often hear terms thrown around when discussing intelligent automation, but what do they really mean? Here, I got you. 

Citizen Developer
Business users with varied coding experience who build business-critical applications using no-code or low-code technology

Computer Vision
The capability of software to interpret the world visually through cameras, video, and images

Deep Learning
A subset of machine-learning that teaches computers to imitate the way humans think and learn

Exceptions
A value not confidently captured by the model

F1-Score
A mean of the precision and recall

Field
A value defined by the customer that they want to extract 

GPU (Graphic Processing Unit)
A computer chip with high computational power designed to process large amounts of data. Recommended for AI software. 

Human-in-the-Loop
Having a human involved in the model training process (i.e. enabling them to make changes/corrections to the training data to enhance model accuracy) 

Machine Learning
How we “teach” a computer model to make predictions and draw conclusions from data

Natural Language Processing (NLP)
The capability for a computer to interpret written or spoken language

Optical Character Recognition (OCR)
Converting handwritten, typed, or scanned text into machine-readable text 

Precision
The percentage of predicted values that match the actual values

Recall
The percentage of correct fields identified by the model

Semi-Structured Data
Data that is partially defined and searchable 

Semi-supervised learning 
A training method where the training set is composed of both “known” (labelled) and “unknown” (unlabelled) data

Straight-through processing
End-to-end automation of a process; ideally, requires no human intervention. 

Structured Data
Data that is clearly defined, searchable, and can be organized into a clear, tabular format

Supervised learning
A training method using “known” (labelled) data to train a model; the values we train the model to predict are pre-defined, or based off known inputs in the data

Transfer Learning
A machine-learning technique that focuses on storing knowledge gained while solving one problem and applying it to a different, but related, problem

Unsupervised learning
A training method using “unknown” (unlabelled) data to train a model; the model is analyzing and organizing the data into meaningful groups based on self-similar characteristics

Unstructured Data
Data that comes in all forms and cannot be easily processed by machines

Leave a Reply