Difficulty: Intermediate
Optimal control for linear systems with quadratic cost.
Minimum relative to nearby points.
Penalizes confident wrong predictions heavily; standard for classification and language modeling.
Raw model outputs before converting to probabilities; manipulated during decoding and calibration.
PEFT method injecting trainable low-rank matrices into layers, enabling efficient fine-tuning.
A function measuring prediction error (and sometimes calibration), guiding gradient-based optimization.
The shape of the loss function over parameter space.
An RNN variant using gates to mitigate vanishing gradients and capture longer context.
Stability proven via monotonic decrease of Lyapunov function.
A subfield of AI where models learn patterns from data to make predictions or decisions, improving with experience rather than explicit rule-coding.
Bayesian parameter estimation using the mode of the posterior distribution.
Mechanics of price formation.
Formal framework for sequential decision-making under uncertainty.
Predicts masked tokens in a sequence, enabling bidirectional context; often used for embeddings rather than generation.
Estimating parameters by maximizing likelihood of observed data.
Average of squared residuals; common regression objective.
AI applied to X-rays, CT, MRI, ultrasound, pathology slides.
Mechanisms for retaining context across turns/sessions: scratchpads, vector memories, structured stores.
Extending agents with long-term memory stores.
GNN framework where nodes iteratively exchange and aggregate messages from neighbors.
Methods that learn training procedures or initializations so models can adapt quickly to new tasks with little data.
Routes inputs to subsets of parameters for scalable capacity.
Practices for operationalizing ML: versioning, CI/CD, monitoring, retraining, and reliable production management.
A parameterized mapping from inputs to outputs; includes architecture + learned parameters.
Standardized documentation describing intended use, performance, limitations, data, and ethical considerations.
Requirement to reveal AI usage in legal decisions.
Required descriptions of model behavior and limits.
Policies and practices for approving, monitoring, auditing, and documenting models in production.
Central catalog of deployed and experimental models.
Inferring sensitive features of training data.