A
Active Learning
A machine learning approach where the algorithm can query a user for labels.
Adversarial Training
A technique to make models more robust by training on adversarial examples.
Algorithm
A set of rules or instructions given to an AI system to help it learn or make decisions.
Annotation
The process of labelling data to make it usable for machine learning models.
Attention Mechanism
A technique that mimics cognitive attention, focusing on specific parts of the input.
AyaGrow
An AI-powered agricultural intelligence platform designed for African farming conditions.
AyaSpeech
An AI-powered speech technology solution specialising in African languages.
B
Backpropagation
An algorithm used to calculate gradients in neural networks.
Batch Normalization
A technique to improve the stability and performance of neural networks.
Bayesian Optimization
A strategy for optimising black-box functions.
Bias
Systematic error introduced by a model or measurement.
Bounding Box
In object detection, a rectangle enclosing an object in an image.
C
Classification
Categorising data into predefined classes.
Clustering
Grouping similar data points together.
Computer Vision
AI field focuses on how computers gain understanding from digital images or videos.
Convolutional Neural Network (CNN)
A deep learning architecture particularly effective for image processing.
Cross-validation
A resampling procedure used to evaluate machine learning models.
D
Data Acquisition
The process of gathering data from various sources for use in AI models.
Data Augmentation
Techniques to increase the amount of training data.
Data Cleaning
The process of fixing or removing incorrect, corrupted, incorrectly formatted, duplicate, or incomplete data.
Data Labelling
The process of adding meaningful tags to data.
Deep Learning
A subset of machine learning using neural networks with multiple layers.
Dimensionality Reduction
Reducing the number of random variables under consideration.
E
Embedding
A representation of discrete variables as vectors in a continuous vector space.
Ensemble Methods
Grouping similar data points together.
Epoch
One complete pass through the entire training dataset.
Explainable AI (XAI)
AI systems whose actions can be easily understood by humans
F
Feature Engineering
The process of using domain knowledge to extract features from raw data.
Few-Shot Learning
Learning a new task from only a few examples.
Fine-tuning
Adapting a pre-trained model to a new, similar task.
G
Generative Adversarial Network (GAN)
A framework for estimating generative models via an adversarial process.
Gradient Descent
An optimisation algorithm for finding the minimum of a function
H
Hyperparameter
A parameter whose value is set before the learning process begins.
Hyperparameter Tuning
The process of optimising hyperparameters in machine learning models.
I
Image Segmentation
The process of partitioning a digital image into multiple segments.
Imbalanced Data
A situation where the classes in a classification problem are not represented equally.
Inference
Using a trained model to make predictions.
K
K-Means Clustering
An unsupervised learning algorithm for clustering.
Kernel Trick
A method for using a linear classifier to solve a non-linear problem.
Keypoint Annotation
Marking specific points of interest in images or videos.
L
LiDAR (Light Detection and Ranging)
A remote sensing method that uses light in the form of a pulsed laser to measure distances.
LSTM (Long Short-Term Memory)
A type of RNN capable of learning long-term dependencies.
Loss Function
A method of evaluating how well an algorithm models the given data.
M
Machine Learning:
A subset of AI that enables systems to learn from data.
Model Compression
Techniques to reduce the size of a model without significantly affecting its performance.
Multi-task Learning
An approach to inductive transfer that improves generalisation by using the domain information contained in the training signals of related tasks.
N
Natural Language Processing (NLP)
A subfield of AI concerned with the interactions between computers and human language.
Neural Architecture Search
The process of automating the design of artificial neural networks.
O
Object Detection
The task of identifying and locating objects in an image or video.
One-Shot Learning
The task of learning information about object categories from one, or only a few, training samples/images.
Overfitting
When a model learns the training data too well, including noise and details irrelevant to the general pattern.
P
Point Cloud
A set of data points in space, often used to represent 3D shapes or scenes.
Polygon Annotation
In image annotation, drawing polygons around objects of interest for more precise object delineation.
Pooling
A technique to reduce the dimensions of feature maps in CNNs.
Pre-training
Training a model on a large dataset before fine-tuning it on a more specific task.
Q
Q-Learning
A model-free reinforcement learning algorithm to learn the value of an action in a particular state.
R
RAG (Retrieval-Augmented Generation)
A technique that combines retrieval-based and generation-based approaches in natural language processing tasks.
Recurrent Neural Network (RNN)
A class of neural networks where connections between nodes form a directed graph along a temporal sequence.
Regularization
Techniques used to reduce the error by fitting a function appropriately on the given training set and avoid overfitting.
Reinforcement Learning
An area of machine learning concerned with how software agents ought to take actions in an environment to maximise cumulative reward.
RLHF (Reinforcement Learning from Human Feedback)
A machine learning technique that incorporates human feedback to guide the learning process of AI models.
S
Satellite Imagery
Using satellite-captured images for data analysis and model training.
Self-Supervised Learning
A type of learning where the data provides the supervision.
Semantic Segmentation
The process of assigning a class label to each pixel in an image.
Sentiment Analysis
The use of NLP to systematically identify, extract, quantify, and study affective states and subjective information.
Speech Recognition
The ability of a machine or program to identify words and phrases in spoken language and convert them to text.
Supervised Learning
A type of machine learning task that uses labelled training data.
Support Vector Machine (SVM)
A supervised learning model used for classification and regression tasks.
T
Transfer Learning
A technique where a model developed for a task is reused as the starting point for a model on a second task.
Transformer
A deep learning model that adopts the mechanism of self-attention.
U
Unsupervised Learning
A type of machine learning that looks for previously undetected patterns in a dataset without pre-existing labels.
V
Validation Set
A set of data used to tune the hyperparameters of a model.
Variational Autoencoder (VAE)
A type of generative model that learns to encode and decode data.
W
Web Scraping
Automatically extracting data from websites.
Word Embedding
Representations of words that allow words with similar meaning to have a similar representation.
Y
Yield Prediction
Using AI to forecast crop yields based on various data inputs
Z
Zero-Shot Learning
The task of learning to recognise new classes not seen during training.
