AI Reflection, Project Cycle and Ethics

Vocabulary practice cards covering the foundations of Artificial Intelligence, the stages of the AI Project Cycle, evaluation metrics like the confusion matrix, and the principles of AI Ethics.

Artificial Intelligence

A machine that mimics human intelligence, can solve real-world problems, improves on its own from past experiences, and can predict and make decisions on its own.

Computer Vision (CV)

An AI domain that works with videos and images enabling machines to interpret and understand visual information.

Natural Language Processing (NLP)

An AI domain focused on textual data enabling machines to comprehend, generate, and manipulate human language.

Statistical Data (AI Domain)

An AI domain that refers to statistical techniques used to analyse, interpret, and draw insights from numerical/tabular data.

Problem Scoping

The first step of the AI project cycle, which involves defining the problem which you wish to solve with the help of your project.

Data Acquisition

The stage of acquiring accurate and reliable data from different sources to ensure the efficiency of the AI system.

Data Exploration

The stage of interpreting useful information out of acquired data to identify patterns and trends related to factors like infestations or crop yields.

Modelling

The study and selection of different AI enabled systems or algorithms that match the project requirements.

Evaluation

The process of testing the developed system in as many ways as possible to choose the model that gives the best results.

Deployment

The final stage where the solution based on the selected model is deployed, such as installing a management system app on a user's phone.

4Ws Canvas

A tool used in Problem Scoping that identifies the Who (stakeholders), What (issue or problem), Where (context or situation), and Why (benefit of the solution).

Data features

The type of data you want to collect; it is a measurable property or characteristic of the data object being stored.

Categorical Data

Qualitative data that represents a type or category rather than a numerical value.

Numeric Data

Quantitative data that represents measurable amounts or numbers.

System Map

A graphical representation of the components of a system and their interrelationships.

Nodes

Entities that define a system within a System Map.

Boundaries

The scope of the system as defined in a System Map.

Machine Learning (ML)

A subset of AI that enables machines to improve at tasks with experience by learning from mistakes.

Deep Learning (DL)

A subset of AI that enables software to train itself to perform tasks.

Rule Based Approach

An AI modelling approach where the relationship or patterns in data are defined by the developer through specific instructions.

Learning Based Approach

An AI modelling approach where patterns and trends are not defined by the developer, but are figured out by the machine from random data.

Overfitting

A situation where the same data is used for building and testing, causing the model to always predict labels correctly regardless of context.

True Positive (TP)

A case where the prediction and reality match (True) and the prediction is Positive.

True Negative (TN)

A case where the prediction and reality match (True) and the prediction is Negative.

False Positive (FP)

A case where the prediction and reality do not match (False) and the prediction is Positive.

False Negative (FN)

A case where the prediction and reality do not match (False) and the prediction is Negative.

Morals

The beliefs dictated by society which are not fixed and can be different for different societies.

Ethics

The guiding principles a person chooses for their life to decide what is good or bad.

Bias

Partiality or preference for one over another that often comes from training data and appears in AI results.

Inclusion

An AI principle regarding whether the system leaves out any groups and if benefits are shared equally regardless of status.