AI Terms & Definitions

A concept from educational psychology applied to AI, referring to tasks an AI can perform with guidance but not independently.

The ability of a model to correctly perform tasks it has not explicitly been trained to do, demonstrating generalization.

Referring to diligent, hard work, often of a nature that is repetitive or requires a high level of effort and reliability.

A technique in NLP where words are represented as vectors in a high-dimensional space, capturing semantic similarity.

AI designed and trained for a specific task, lacking the general cognitive abilities of human intelligence.

A specialized database for storing and retrieving vector representations of data, facilitating similarity searches.

The encoding of words or phrases as numerical vectors, enabling mathematical operations and comparisons by AI models.

Different expressions or phrasings that convey the same intent or meaning, important in understanding natural language variability.

In machine learning, the amount by which the model's predictions vary from the average prediction, reflecting sensitivity to training data.

Data set aside from the training dataset to tune model parameters and prevent overfitting, ensuring the model's generalizability.

The process of evaluating a model's performance with a separate portion of the data not used in training, to gauge its accuracy.

The means by which a human interacts with a computer, application, or machine, often focusing on ease of use.

A technique in generative AI where multiple outputs are generated and the best one is selected based on certain criteria.

A type of machine learning where models learn patterns from unlabeled data, without explicit instructions.

A test of a machine's ability to exhibit intelligent behavior indistinguishable from that of a human.

A class of deep learning models that have revolutionized the field of natural language processing (NLP).

The component of a transformer model responsible for generating output sequences based on encoded information.

A model architecture that uses self-attention mechanisms to improve performance on tasks involving sequential data.

Leveraging knowledge gained while solving one problem to solve a different but related problem in machine learning.

The dataset used specifically for training a machine learning model, containing examples for learning patterns and behaviors.

The process of teaching a machine learning model to make predictions or decisions, typically by exposing it to a large dataset.

A statistical model to discover abstract topics within a collection of documents, aiding in content organization and discovery.

The smallest unit of processing in NLP, which could be a word, part of a word, or a character, depending on the model.

The task of assigning predefined categories to text, used in applications like spam detection and sentiment analysis.

A dataset used to evaluate the performance of a machine learning model after training, separate from training data.

Internal cues or instructions that guide the behavior of an AI model, influencing how it processes and responds to input.

A machine learning approach where models are trained on labeled data, learning to predict outcomes from inputs.

Predefined messages or prompts used in conversational AI systems to guide user interactions.

The process of refining a model's performance on specific tasks by training it further with labeled data.

AI with the ability to understand, learn, and apply knowledge in ways indistinguishable from human intelligence.

Models that transform sequences from one domain to another, commonly used in translation and summarization.

The process where AI models produce a sequence of items, such as words in text generation, based on learned patterns.

The computational task of identifying and categorizing opinions expressed in text to determine the writer's attitude.

The measure of how much two pieces of text are related in terms of meaning, used in various NLP tasks.

Search technology that understands the context and intent behind a user's query to generate more relevant results.

The process of adding semantic metadata to content, making it easier for AI to understand and process information.

A mechanism that allows models to weigh the importance of different parts of the input data relative to each other.

Observations that as AI models increase in size, their performance improves according to predictable patterns.

A testing environment that isolates untested code changes and experimentation without affecting the production environment.

Models that evaluate potential actions or responses in reinforcement learning to guide learning towards desired outcomes.

A model that retrieves relevant information from a large dataset to support decision-making or responses.

Combining retrieval of relevant information with generative models to produce informed responses.

An evaluation of how well an AI system's responses meet the criteria of relevance, coherence, and accuracy.

Training approach where models are refined based on feedback from human evaluators.

A type of machine learning where an agent learns to make decisions by taking actions in an environment to achieve rewards.

Techniques used to prevent overfitting by penalizing complex models during the training process.

A type of neural network well-suited for processing sequences of data, like text or time series.

A request for information or action made to a database, search engine, or AI model.

A system that automatically answers questions posed by humans in a natural language.

A high-level programming language known for its clear syntax and readability, widely used in AI development.

A reinforcement learning algorithm that balances exploration and exploitation in policy learning.

A technique used to influence or manipulate the behavior of AI systems through specially crafted inputs.

The art of crafting prompts to effectively communicate with and elicit desired responses from AI models.

A text input given to an AI model, designed to elicit a specific type of response or output.

The initial training phase where a model learns from a large, general dataset before task-specific training.

A model that makes predictions about unknown future events based on patterns found in historical data.

The use of data, statistical algorithms, and machine learning techniques to identify the likelihood of future outcomes.

Additional software components that extend or enhance the functionality of an AI system or application.

The automated recognition of patterns and regularities in data using machine learning algorithms.

The process of marking up a word in a text as corresponding to a particular part of speech.

A variable in a model that is learned from the training data and determines the model's output.

A technique to discourage repetitive or overly similar responses in generative AI models.

A modeling error in machine learning where a model learns the detail and noise in the training data too well.

An AI research lab focusing on developing and promoting friendly AI for the benefit of humanity.

A model training approach where the model updates continuously as new data arrives.

The ability of a model to learn information from a single example or a few examples.

Learning techniques where the model learns from one or a few examples, respectively.

Learning optimal actions from a fixed dataset without further interaction with the environment.

A series of algorithms that mimic the operations of a human brain to recognize relationships in data.

The ability of AI to understand and interpret human language as it is spoken or written.

The field of AI focused on the interaction between computers and humans through natural language.

Generating coherent and contextually relevant text from structured data using AI.

The process of identifying and classifying key information (entities) in text into predefined categories.

Conversations where participants exchange multiple sequences of messages, requiring context understanding.

Training an AI model on multiple tasks simultaneously, leveraging commonalities across tasks.

AI systems that can process and interpret multiple types of data, such as text, images, and sound.

Tools designed to monitor and manage the behavior of AI systems, ensuring they adhere to guidelines.

A comprehensive document providing essential information about a machine learning model's purpose and performance.

The specific layout and structure of a machine learning model, including its layers and nodes.

A mathematical representation of a real-world process, trained to perform specific tasks using data.

The largest amount of text or data that a model can generate in response to a single prompt.

A training technique where some words in the input are hidden, and the model predicts them.

A mathematical framework for modeling decision-making in situations with random outcomes.

The use of software to translate text or speech from one language to another automatically.

The science of getting computers to act without being explicitly programmed, through learning.

Broad term encompassing the capabilities of machines to learn from data and perform tasks.

A technique for fine-tuning large models in a memory and computationally efficient manner.

The process of adding metadata regarding linguistic information to text, aiding in its analysis.

Hidden or unobservable variables inferred from observable data in machine learning models.

An extensive model trained on vast amounts of text data, capable of understanding and generating text.

AI that understands, interprets, and generates human language based on statistical probabilities.

A tag or annotation applied to a piece of data, indicating the correct output for supervised learning.

The method by which AI systems model, store, and retrieve knowledge to solve complex tasks.

A centralized repository of information, used in AI to provide answers or contextual information.

The probability of two events happening at the same time in a probabilistic model.

The underlying purpose or goal that a user aims to achieve through a query or statement.

A variant of GPT trained to follow instructions in prompts and generate more specific responses.

The process of automatically extracting structured information from unstructured text data.

The phase where a trained model is used to make predictions or decisions based on new, unseen data.

A parameter whose value is set before the learning process begins, influencing the training phase.

Problem-solving approaches that use practical methods or various shortcuts to produce solutions.

When AI generates information that is not grounded in reality, often due to training data issues.

Optimization algorithms that make the locally optimal choice at each step to find a global optimum.

The third iteration of OpenAI's generative model known for its advanced text generation capabilities.

In GANs, the component that creates data aiming to mimic real data as closely as possible.

A type of AI model specializing in generating coherent and contextually relevant text.

A type of AI model that can generate new data instances similar to the training data.

AI systems capable of generating new, original content or data that mimics real-world examples.

A framework for training generative models through a competitive process between networks.

The process of producing new content, such as text or images, based on learned patterns and data.

Artificial intelligence that exhibits cognitive functions across a wide range of tasks and domains.

A large, versatile AI model trained on a broad dataset, capable of performing multiple tasks.

A logical reasoning method that starts with known facts and applies rules to reach new conclusions.

The process of adjusting a pre-trained model to perform well on a specific task or dataset.

The capability to precisely adjust the output or behavior of an AI model based on specific criteria.

The ability of a model to learn and generalize from a very small number of examples.

Identifying and isolating useful information from data to improve model training and performance.

Creating summaries by extracting key sentences or fragments directly from the source text.

AI systems designed to provide insights into their decision-making processes for transparency.

Quantitative measures used to assess the performance and effectiveness of AI models.

A framework for assessing and guiding the ethical development and deployment of AI systems.

Identifying and classifying named entities in text into predefined categories.

The process of labeling text with information about entities, enhancing data structure.

Specific, identifiable elements in text, such as names, places, dates, often extracted by AI.

The application of artificial intelligence technologies to improve business processes and outcomes.

A component of a model that processes and transforms input data into a usable format.

Dense vector representations of words or phrases capturing semantic meaning for AI processing.

A method where AI model training is spread across multiple computers or servers.

The component of a generative adversarial network that distinguishes real data from fake.

AI technologies designed to converse with humans using natural language processing.

The process of making an AI model available for use in real-world applications or systems.

Analyzing the grammatical structure of a sentence to understand relationships between words.

A subset of machine learning involving neural networks with many layers to analyze data.

Guidelines that dictate how a language model translates its internal representations to output.

A collection of data specifically prepared and structured for training or testing AI models.

An interdisciplinary field that uses scientific methods to extract knowledge from data.

Measures and practices to ensure that personal or sensitive data is not misused or disclosed.

The practice of examining large databases to generate new information and find hidden patterns.

A technique for increasing the amount of training data by adding slightly modified copies.

Integrating artificial intelligence into customer relationship management to enhance interactions.

A large collection of texts used for compiling data and training machine learning models.

The task in NLP of determining which words refer to the same entity in a text.

The range of past input that a model can consider when generating a response or prediction.

Representations of words or phrases that take into account the context in which they appear.

A branch of artificial intelligence focused on understanding the algorithms that drive learning.

The output produced by AI in response to a given input or prompt, completing the thought process.

AI systems designed to mimic human brain functioning, aiming for natural, human-like interaction.

An AI developed by OpenAI that can generate human-like text responses based on provided prompts.

Computer programs designed to simulate conversation with human users, often over the internet.

A prompting strategy that encourages AI to break down complex problems into manageable steps.

A rectangular border used in visual processing to define the location of objects within images.

Extremely large data sets analyzed computationally to reveal patterns, trends, and associations.

Preconceived notions or predispositions in AI models that can affect decision-making and fairness.

A search algorithm that efficiently finds the most likely sequences of outcomes in models.

A strategy for balancing the exploration of new choices and the exploitation of known rewards.

A reasoning method that starts with the end goal and works backward to determine the solution path.

A method used in training artificial neural networks, adjusting weights based on error rates.

Models that use previous time points to predict future values, common in time series forecasting.

A statistical model that predicts future behavior based on past outcomes in time series data.

Machines or systems capable of performing tasks and making decisions without human intervention.

Enhancing human decision-making with AI capabilities, focusing on collaboration between humans and AI.

In AI, a technique that helps models focus on relevant parts of the input data, improving relevance.

A mechanism in AI that allows models to weigh the importance of different pieces of information.

Computing systems vaguely inspired by the biological neural networks in human brains.

Interfaces that allow different software applications to communicate and work together.

The process of ensuring AI behaviors and outputs adhere to human ethical standards and intentions.

A set of mathematical instructions or rules that a computer follows to perform a specific task efficiently.

Specialists who refine and enhance AI models by providing feedback on outputs and guiding learning.

The simulation of human intelligence processes by machines, particularly computer systems.

AI entities capable of autonomously performing tasks across various domains, akin to digital assistants.

Improves AI robustness by training with deliberately challenging inputs to enhance model accuracy.