Artificial Intelligence

A capability of language models that allows them to generate structured outputs requesting the execution of predefined functions or tools. Function calling enables LLMs to interact with external APIs,...

The use of Graphics Processing Units for general-purpose computing tasks, particularly matrix operations central to machine learning. GPUs can process thousands of operations in parallel, making them...

A type of artificial intelligence that can create new content such as text, images, music, code, and video by learning patterns from existing data. Models like GPT, DALL-E, and Stable Diffusion are pr...

A framework where two neural networks compete: a generator creates synthetic data, and a discriminator evaluates its authenticity. Through this adversarial process, GANs can produce highly realistic i...

A family of autoregressive language models developed by OpenAI that use transformer architecture to generate human-like text. GPT models are pre-trained on large text corpora using next-token predicti...

An optimization algorithm used to minimize the loss function in machine learning models by iteratively adjusting parameters in the direction of the steepest descent. Variants include stochastic gradie...

The process of connecting AI model outputs to verifiable, real-world information sources. Grounding reduces hallucinations by ensuring that generated content is based on factual data, retrieved docume...

Safety mechanisms and constraints implemented to control and limit AI system behavior. Guardrails include input/output filters, content policies, rate limits, and structured validation that prevent AI...

A phenomenon where an AI model generates information that sounds plausible but is factually incorrect or fabricated. Hallucinations occur because language models generate text based on statistical pat...

A configuration variable set before the training process that controls the learning behavior, such as learning rate, batch size, number of epochs, and network architecture. Unlike model parameters, hy...

The process of using a trained machine learning model to make predictions or generate outputs on new, unseen data. Inference is the deployment phase of AI, as opposed to the training phase, and is opt...

A type of AI model trained on vast amounts of text data that can understand, generate, and reason about human language. LLMs like GPT-4, Claude, and Llama use transformer architectures with billions o...

A specialized type of recurrent neural network designed to learn long-range dependencies in sequential data. LSTMs use gate mechanisms (input, forget, output) to control information flow, solving the...

A mathematical function that measures the difference between the predicted output and the actual target value. The loss function guides the training process by quantifying how well or poorly a model i...

A parameter-efficient fine-tuning technique that freezes the pre-trained model weights and injects trainable low-rank decomposition matrices into transformer layers. LoRA drastically reduces the numbe...

A set of practices that combines machine learning, DevOps, and data engineering to deploy and maintain ML models in production reliably and efficiently. MLOps covers model versioning, experiment track...

A branch of artificial intelligence that enables systems to learn and improve from experience without being explicitly programmed. Machine learning algorithms use statistical techniques to identify pa...

A neural network architecture that divides the model into multiple specialized sub-networks (experts) and uses a gating mechanism to route each input to the most relevant experts. MoE enables scaling...

An open standard that enables AI models to securely connect to external data sources and tools through a unified protocol. MCP provides a standardized way for AI applications to access context from da...

A technique for compressing a large, complex model (teacher) into a smaller, more efficient model (student) that approximates the teacher's performance. The student learns from the teacher's output pr...