AI & Robotics Glossary

A mathematical function applied to a neuron's output that introduces non-linearity into the network, enabling it to learn complex patterns. Common examples include ReLU, sigmoid, and tanh.

A component in a robot that converts energy into motion. Actuators are responsible for moving and controlling the robot's limbs, joints, and mechanisms.

The moral principles and guidelines governing the development and use of AI systems, addressing issues like bias, privacy, transparency, accountability, and societal impact.

The simulation of human intelligence processes by computer systems, including learning, reasoning, problem-solving, perception, and language understanding.

A technique that allows models to focus on relevant parts of the input when producing output. It assigns different weights to different parts of the input sequence.

The algorithm used to calculate gradients in neural networks by propagating errors backward from the output layer to earlier layers, enabling efficient training of deep networks.

The number of training examples used in one iteration of model training. Larger batches provide more stable gradients but require more memory.

A field of AI that trains computers to interpret and understand the visual world using digital images and deep learning models to accurately identify and classify objects.

A deep learning architecture specifically designed for processing structured grid data like images. It uses convolutional layers to automatically learn spatial hierarchies of features.

The process by which quantum systems lose their quantum properties (superposition and entanglement) due to interaction with their environment. Decoherence is the main obstacle to building practical quantum computers, as it causes errors in quantum calculations.

A subset of machine learning based on artificial neural networks with multiple layers (deep neural networks). It excels at finding patterns in unstructured data like images, sound, and text.

The number of independent movements or parameters that define a robot's configuration. Each joint typically adds one degree of freedom to a robotic system.

AI algorithms processed locally on hardware devices (edge devices) rather than in centralized cloud servers, enabling faster response times and better privacy.

A learned representation that maps discrete items (like words or categories) to continuous vectors in a lower-dimensional space, capturing semantic relationships.

The device at the end of a robotic arm that interacts with the environment. It can be a gripper, tool, sensor, or any other device designed for a specific task.

One complete pass through the entire training dataset during the training process. Multiple epochs are typically needed for a model to learn effectively.

The process of further training a pre-trained model on a specific dataset to adapt it for a particular task or domain, improving its performance for specialized applications.

A deep learning architecture consisting of two neural networks (generator and discriminator) that compete against each other to generate new, synthetic data that resembles training data.

A family of large language models developed by OpenAI that use the transformer architecture. GPT models are pre-trained on large text datasets and can generate coherent, contextual text.

An optimization algorithm used to minimize the loss function in machine learning by iteratively adjusting model parameters in the direction of steepest descent.

When an AI model generates information that sounds plausible but is factually incorrect, made up, or not grounded in its training data. A significant challenge in LLM reliability.

A robot with a body shape designed to resemble the human body, typically with a head, torso, two arms, and two legs. Designed to interact naturally with human environments and tools.

Configuration settings set before training that control the learning process, such as learning rate, batch size, and number of layers. Unlike parameters, they're not learned from data.

The process of using a trained machine learning model to make predictions on new, unseen data. It's the deployment phase after training is complete.

The mathematical process of calculating the joint angles needed to place a robot's end effector at a desired position and orientation in space.

A type of AI model trained on massive amounts of text data to understand and generate human-like text. LLMs can perform various language tasks like translation, summarization, and conversation.

A special type of RNN architecture designed to learn long-term dependencies. It uses gates to control information flow, solving the vanishing gradient problem in standard RNNs.

A function that measures how well a model's predictions match the actual target values. The goal of training is to minimize this loss.

A subset of AI that enables systems to automatically learn and improve from experience without being explicitly programmed. It focuses on developing algorithms that can access data and use it to learn for themselves.

AI systems that can process and understand multiple types of data inputs simultaneously, such as text, images, audio, and video, to provide more comprehensive understanding.

A branch of AI that helps computers understand, interpret, and generate human language. It combines computational linguistics with machine learning to process and analyze large amounts of natural language data.

A computing system inspired by biological neural networks in the human brain. It consists of interconnected nodes (neurons) organized in layers that process information using connectionist approaches.

When a model learns the training data too well, including its noise and outliers, resulting in poor performance on new, unseen data. The model memorizes rather than generalizes.

The internal variables of a machine learning model that are learned from training data. The number of parameters often indicates model complexity and capability.

The practice of designing and optimizing input prompts to get desired outputs from AI language models. It involves crafting clear, specific instructions to guide the model's responses.

An algorithm designed to run on quantum computers, leveraging quantum mechanical phenomena like superposition and entanglement to solve problems faster than classical algorithms. Famous examples include Shor's algorithm for factoring and Grover's algorithm for searching.

A quantum computing approach specifically designed for optimization problems. Unlike gate-based quantum computers, quantum annealers use quantum fluctuations to find the lowest energy state of a system, which represents the optimal solution.

A quantum phenomenon where two or more particles become correlated in such a way that the quantum state of each particle cannot be described independently. Measuring one particle instantly affects its entangled partner, regardless of distance.

Techniques used to protect quantum information from errors due to decoherence and other quantum noise. Unlike classical error correction, quantum error correction must preserve superposition and entanglement while detecting and correcting errors.

The basic building block of quantum circuits, analogous to classical logic gates. Quantum gates manipulate qubits through unitary operations, changing their quantum states. Common gates include Hadamard, CNOT, and Pauli gates.

The integration of quantum computing with machine learning algorithms. QML aims to speed up ML tasks by using quantum properties to process data more efficiently, potentially enabling breakthroughs in pattern recognition, optimization, and AI training.

A fundamental principle of quantum mechanics where a quantum system can exist in multiple states at the same time until it is measured. This property allows quantum computers to explore many possible solutions simultaneously.

The point at which a quantum computer can perform a calculation that would be practically impossible for any classical computer to complete in a reasonable time. Google claimed to achieve this milestone in 2019 with their Sycamore processor.

The fundamental unit of quantum information, analogous to a classical bit. Unlike classical bits that can only be 0 or 1, qubits can exist in a superposition of both states simultaneously, enabling quantum computers to process vast amounts of information in parallel.

A technique that enhances LLM responses by first retrieving relevant information from external knowledge bases, then using that context to generate more accurate, up-to-date answers.

A neural network architecture designed for sequential data where connections between nodes form directed cycles, allowing information to persist and enabling the network to learn from sequences.

A type of machine learning where an agent learns to make decisions by performing actions in an environment to maximize cumulative reward. It learns through trial and error.

An open-source framework for robot software development. It provides tools, libraries, and conventions to simplify creating complex robot behaviors across various robotic platforms.

A computational technique that allows robots to build a map of an unknown environment while simultaneously keeping track of their location within it.

A type of machine learning where the model is trained on labeled data. The algorithm learns from input-output pairs to make predictions on new, unseen data.

The process of breaking down text into smaller units called tokens, which can be words, subwords, or characters. It's a crucial preprocessing step for NLP models.

The dataset used to teach a machine learning model to make predictions or decisions. The quality and quantity of training data directly impacts model performance.

A machine learning technique where a model trained on one task is repurposed as the starting point for a model on a different but related task, saving time and computational resources.

A deep learning architecture that uses self-attention mechanisms to process sequential data in parallel. It revolutionized NLP and is the foundation of models like GPT and BERT.

A type of machine learning where the model learns patterns from unlabeled data without predefined categories. It discovers hidden structures in data.