AI alignment refers to the field of research concerned with ensuring that AI systems behave per human intentions & values
Artificial General Intelligence are hypothetical machines with human-like learning and thinking abilities to perform any intellectual task.
AI is the simulation of human intelligence in machines to perform tasks that require human cognitive abilities.
ANNs mimic the human brain, processing data through interconnected neurons for adaptive learning in machine learning.
ASI is a theoretical concept imagining AI surpassing human intellect, awaiting realisation with AGI as a precursor.
An autoregressive model predicts future sequence values based on its past values using statistical techniques.
Bayesian networks are graphical models that help understand and reason about complex systems with uncertainty using directed graphs.
Big data refers to extremely large and diverse collections of data. They are so massive that data management systems can’t handle them.
Chatbot is a computer programme that simulates conversation with humans, commonly used to provide customer service or answer FAQs.
Composite AI is an approach to AI that combines multiple AI techniques and technologies to create a more comprehensive and capable system.
Conditional generation in AI and ML is the process of creating outputs based on specific conditions or constraints once inputs are given.
Conversational AI is a branch of AI that allows machines to simulate conversation with humans.
CNN is a deep learning algorithm tailored for analysing visual data such as images and videos.
DBN is an advanced neural network adept at extracting complex patterns from large datasets in unsupervised learning.
Emotion AI is a developing field of AI that deals with machines’ ability to recognise and understand human emotions
An encoder-decoder architecture in machine learning efficiently translates one sequence data form to another.
Ethical AI involves creating and deploying AI systems while prioritising ethical principles in their development and use.
Explainable AI (XAI) is a field of AI that focusses on developing techniques to make AI models more understandable to humans.
Fuzzy logic is a thinking process that goes beyond traditional “true or false” logic in processing information.
Generative AI (GenAI) is subfield of Artificial Intelligence for creating content based on a given prompt for text, image, audio or video.
Generative Adversarial Network is an ML model that uses two competing neural networks to generate indistinguishable data.
A generative model is an AI model that learns data patterns to generate new data similar to its training data.
The generative pre-trained transformer is an LLM that is used in generative artificial intelligence (GenAI).
In machine learning, a hierarchical model is an approach that organises data and learning processes into layered structures.
Hybrid AI is essentially a teamwork approach between two different AI techniques – machine learning and symbolic AI.
LLMs is an advanced language model capable of near-human language understanding and generation in AI applications.
A latent space is a kind of hidden space, often with many dimensions, that captures the important features of a set of data.
ML is a subset of artificial intelligence (AI) that allows computers to learn without being explicitly programmed.
Markov Chain Monte Carlo (MCMC) is used in statistics & various scientific fields to sample from complex probability distributions.
Natural Language Generation, an AI process, enables computers to generate human-like text in response to data or information inputs.
Natural language understanding is an AI branch for computers to grasp human language and blend CS, linguistics, and psychology.
A Neural Radiance Field is a recent advancement in deep learning that reconstructs a 3D scene from a collection of 2D images
Overfitting is a common problem in machine learning that occurs when a model becomes too aligned with the training data.
Predictive analytics is a powerful tool that uses data to forecast future outcomes and trends.
A probabilistic model predicts future events by considering uncertainty and expressing possibilities through probabilities.
A probability density function (PDF) describes the likelihood of different outcomes for a continuous random variable.
A quantum generative model, an ML algorithm, uses the principles of quantum mechanics to generate complex data distributions.
RNNs are artificial neural networks designed to handle sequential data like text, speech or financial records.
Reinforcement learning, a subfield of ML, enables intelligent agents to learn optimal behaviour by rewarding and punishing.
Responsible AI is a comprehensive approach to AI that considers the ethical, social and legal implications throughout the entire AI lifecycle
Semi-supervised learning combines the strengths of labelled data and unlabelled data to create effective learning models.
Supervised learning in ML trains algorithms with labelled data, where each data point has predefined outputs, guiding the learning process.
A technological singularity is a hypothetical event where AI surpasses human intelligence, resulting in profound consequences for humanity.
Training data is a collection of examples that the model learns from to identify patterns and make predictions.
A type of neural network architecture that has revolutionised natural language processing in recent years.
Underfitting refers to a scenario in ML where a model fails to capture the underlying patterns within the data effectively.
In ML, unsupervised learning algorithms analyse and discover hidden patterns or structures within unlabelled data.
VAEs are an artificial neural network architecture to generate new data which consist of an encoder and decoder.
Zero data retention is deleting data once its primary purpose has been fulfilled, without intentionally storing it for future use.
Zero-shot learning is an ML approach that uses pre-trained models to classify test data from classes not present during training.
