30 AI Terms Every Tester Should Know

Artificial Intelligence
Artificial intelligence refers to non-human programs that can solve sophisticated tasks requiring human intelligence. For example, an AI system that intelligently identifies images or classifies text. Unlike narrow AI that excels at specific tasks, artificial general intelligence would possess the ability to understand, learn, and apply knowledge across different domains similar to human intelligence.

AI System
An AI system is a comprehensive framework that includes the AI model, datasets, algorithms, and computational resources working together to perform specific functions. AI systems can range from simple rule-based programs to complex generative AI systems capable of creating original content.

Narrow AI
Narrow AI (also called weak AI) refers to artificial intelligence that is focused on performing a specific task, such as image recognition or speech recognition. Most current AI applications use narrow AI, which excels at its programmed function but lacks the broad capabilities of human intelligence.

Expert Point of View: AI is really just a study of intelligent agents. These agents are autonomous, perceive and act on their own within an environment, and generally use sensors and effectors to do so. They analyze themselves with respect to error and success and then adapt, possibly in real time, depending on the application” . This supports the idea of AI systems being comprehensive frameworks capable of learning and adapting.

– Tariq King No B.S Guide to AI in Automation Testing

Machine Learning

Machine Learning

Formally, machine learning is a subfield of artificial intelligence.

However, in recent years, some organizations have begun interchangeably using the terms artificial intelligence and machine learning. Machine learning enables computer systems to learn from and make predictions based on data without being explicitly programmed. Different types of machine learning include supervised learning, unsupervised learning, and reinforcement learning.

Machine Learning Model
A machine learning model is a representation of what a machine learning system has learned from the training data. These learning models form the basis for AI to analyze new data and make predictions.

Machine Learning Algorithm
A machine learning algorithm is a specific set of instructions that allow a computer to learn from data. These algorithms form the backbone of machine learning systems and determine how the model learns from input data to generate outputs.

Machine Learning Techniques
Machine learning techniques encompass various approaches to train AI models, including decision trees, random forests, support vector machines, and deep learning, which use artificial neural network architectures inspired by the human brain.

Machine Learning Systems
Machine learning systems are end-to-end platforms that handle data preprocessing, model training, evaluation, and deployment in a streamlined workflow to solve specific computational problems.

Expert Point of View: “Machine learning is taking a bunch of data, looking at the patterns in there, and then making predictions based on that. It’s one of the core pieces of artificial intelligence, alongside computer vision and natural language processing” . This highlights the role of machine learning models in analyzing data and making predictions.”

– Trevor Chandler QA: Masters of AI Neural Networks

Generative AI

Generative AI
Generative AI is a type of AI model that can create new content such as images, text, or music. These AI tools leverage neural networks to produce original outputs based on patterns learned from training data. Generative AI tools like chatbots have transformed how we interact with AI technologies.

Large Language Model
A large language model is a type of AI model trained on vast amounts of text data, enabling it to understand and generate human language with remarkable accuracy. These models power many conversational AI applications and can perform various natural language processing tasks.

Hallucination
Hallucination occurs when an AI model generates outputs that are factually incorrect or have no basis in its training data. This phenomenon is particularly common in generative AI systems and poses challenges for responsible AI development.

Expert Point of View: “One of the challenges with generative AI is ensuring the outputs are accurate. While these models are powerful, they can sometimes produce results that are incorrect or misleading, which is why understanding their limitations is critical” . This directly addresses the issue of hallucination in generative AI systems.”

– Guljeet Nagpaul Revolutionizing Test Automation: AI-Powered Innovations

Neural Network

Neural Network
A neural network is a computational model inspired by the human brain's structure. It consists of interconnected nodes (neurons) that process and transmit information. Neural networks form the foundation of many advanced machine learning techniques, particularly deep learning.

Artificial Neural Network
An artificial neural network is a specific implementation of neural networks in computer science that processes information through layers of interconnected nodes to recognize patterns in data used to train the model.

Deep Learning
Deep learning is a subset of AI that uses multi-layered neural networks to analyze large amounts of data. These complex networks can automatically extract features from data, enabling breakthroughs in computer vision and speech recognition.

Expert Point of View: “Natural language processing refers to code that gives technology the ability to understand the meaning of text, complete with the writer's intent and their sentiments. NLP is the technology behind text summarization, your digital assistant, voice-operated GPS, and, in this case, a customer service chatbot” ‌1‌‌2‌. This directly supports the idea of NLP enabling computers to interpret and generate human language”

– Emily O’Connor from AG24 Session on Testing AI Chatbot Powered By Natural Language Processing

Types of Learning

Supervised Learning
Supervised learning is a type of machine learning where the model learns from labeled training data to make predictions. The AI system is trained using input-output pairs, with the algorithm adjusting until it achieves the desired accuracy.

Unsupervised Learning
Unsupervised learning involves training an AI on unlabeled data, allowing the model to discover patterns and relationships independently. This form of artificial intelligence is particularly useful when working with datasets where the structure isn't immediately apparent.

Reinforcement Learning
Reinforcement learning is a type of machine learning technique where an AI agent learns by interacting with its environment and receiving feedback in the form of rewards or penalties. This approach has been crucial in developing AI that could master complex games and robotics.

Expert Point of View: “Training a neural network is like teaching it to differentiate between cats and dogs. You feed it data, reward it for correct answers, and adjust weights for wrong ones. Over time, it learns to recognize patterns in the data, much like how humans learn through experience” . This highlights the process of training artificial neural networks to recognize patterns.”

– Noemi Ferrera 

Natural Language Processing

Natural Language Processing
Natural language processing (NLP) is a field within artificial intelligence focused on enabling computers to understand, interpret, and generate human language. NLP powers everything from translation services to conversational AI that can engage in human-like dialogue.

Transformer
A transformer is a type of AI model that learns to understand and generate human-like text by analyzing patterns in large amounts of text data. Transformers have revolutionized natural language processing tasks and form the backbone of many large language models.

Free Automation with Playwright with AI Course

Key AI Terms and Concepts

Model
An AI model is a program trained on data to recognize patterns or make decisions without further human intervention. It uses algorithms to process inputs and generate outputs.

Algorithm
An algorithm is a set of instructions or steps that allow a program to perform computation or solve a problem. Machine learning algorithms are sets of instructions that enable a computer system to learn from data.

Model Parameter
Parameters are internal to the model whose value can be estimated or learned from data. For example, weights are the parameters for neural networks.

Model Hyperparameter
A model hyperparameter is a configuration that is external to the model and whose value cannot be estimated from data. For example, the learning rate for training a neural network is a hyperparameter.

Model Artifact
A model artifact is the byproduct created from training the model. The artifacts will be put into the ML pipeline to serve predictions.

Model Inputs
An input is a data point from a dataset that you pass to the model. For example:

• In image classification, an image can be an input
• In reinforcement learning, an input can be a state

Model Outputs
Model output is the prediction or decision made by a machine learning model based on input data. The quality of outputs depends on both the algorithm and the data used to train an AI model.

Dataset
A dataset is a collection of data used for training, validating, and testing AI models. The quality and amount of data in a dataset significantly impact the performance of machine learning models.

Ground Truth
Ground truth data means the actual data used for training, validating, and testing AI/ML models. It is very important for supervised machine learning.

Data Annotation
Annotation is the process of labeling or tagging data, which is then used to train and fine-tune AI models. This data can be in various forms, such as text, images, or audio used in computer vision systems.

Features
A feature is an attribute associated with an input or sample. An input can be composed of multiple features. In feature engineering, two features are commonly used: numerical and categorical.

Compute
Compute refers to the computational resources (processing power) required to train and run AI models. Advanced AI applications often require significant compute resources, especially for training complex neural networks.

Training and Evaluation

Model Training
Model training in machine learning is “teaching” a model to learn patterns and make predictions by feeding it data and adjusting its parameters to optimize performance. It is the key step in machine learning that results in a model ready to be validated, tested, and deployed. AI training typically requires significant computational resources, especially for complex models.

Fine Tuning
Fine-tuning is the process of taking a pre-trained AI model and further training it on a specific, often smaller, dataset to adapt it to particular tasks or requirements. This technique is commonly used when developing AI for specialized applications.

Inference
A model inference pipeline is a program that takes input data and then uses a trained model to make predictions or inferences from the data. It's the process of deploying and using a trained model in a production environment to generate outputs on new, unseen data.

ML Pipeline
A machine learning pipeline is a series of interconnected data processing and modeling steps designed to automate, standardize, and streamline the process of building, training, evaluating, and deploying machine learning models. ML pipelines aim to automate and standardize the machine learning process, making it more efficient and reproducible.

Model Registry
The model registry is a repository of the trained machine learning models, including their versions, metadata, and lineage. It dramatically simplifies the task of tracking models as they move through the ML lifecycle, from training to production deployments.

Batch Size
The batch size is a hyperparameter that defines the number of samples to work through before updating the internal model parameters.

Batch Vs Real-time processing
Batch processing is done offline. It analyzes large historical datasets all at once and allows the machine learning model to make predictions on the output data. Real-time processing, also known as online or stream processing, thrives in fast-paced environments where data is continuously generated and immediate insights are crucial.

Feedback Loop
A feedback loop is the process of leveraging the output of an AI system and corresponding end-user actions in order to retrain and improve models over time.

Join Our Free Private Community

Model Evaluation and Ethics

Model Evaluation
Model evaluation is a process of evaluating model performance across specific use cases. It might also be referred to as the observability of a model's performance.

Model Observability
ML observability is the ability to monitor and understand a model's performance across all stages of the model development cycle.

Accuracy
Accuracy refers to the percentage of correct predictions a model makes, calculated by dividing the number of correct predictions by the total number of predictions.

Precision
Precision shows how often an ML model is correct when predicting the target class.

Recall, or True Positive Rate(TPR)
Recall is a metric that measures how often a machine learning model correctly identifies positive instances (true positives) from all the actual positive samples in the dataset.

F1-Score
The F1 score can be interpreted as a harmonic mean of precision and recall, where an F1 score reaches its best value at 1 and worst score at 0.

Data Drift
Data drift is a change in the model inputs the model is not trained to handle. Detecting and addressing data drift is vital to maintaining ML model reliability in dynamic settings.

Concept Drift
Concept drift is a change in input-output target variables. It means that whatever your model is predicting is changing.

Bias
Bias is a systematic error that occurs when some aspects of a dataset are given more weight and/or representation than others. There are many types of bias, such as historical bias and selection bias. Addressing bias is a critical component of responsible AI efforts.

AI Ethics
AI ethics encompasses the moral principles and values that guide the development and use of artificial intelligence. This includes considerations around fairness, transparency, privacy, and the social impact of AI technologies in the AI landscape.

Computer Vision

Computer Vision
Computer vision is a field of AI that trains computers to interpret and understand visual information from the world. Image recognition systems are a common application of computer vision technology.

Understanding these key terms will enhance your comprehension of AI concepts and provide a solid foundation for navigating the rapidly evolving field of artificial intelligence. As the AI terminology continues to develop, staying informed about different AI applications and technologies becomes increasingly important for professionals across all industries.