Difficulty: Intermediate

412 terms

Feedback Loop Intermediate

Using production outcomes to improve models.

Feedback Loop Collapse Intermediate

Model trained on its own outputs degrades quality.

Few-Shot Learning Intermediate

Achieving task performance by providing a small number of examples inside the prompt without weight updates.

Fine-Tuning Intermediate

Updating a pretrained model’s weights on task-specific data to improve performance or adapt style/behavior.

Fisher Information Intermediate

Measures how much information an observable random variable carries about unknown parameters.

Flat Minimum Intermediate

A wide basin often correlated with better generalization.

Forced Alignment Intermediate

Aligns transcripts with audio timestamps.

Forecasting Intermediate

Predicting future values from past observations.

Fraud Detection Intermediate

Identifying suspicious transactions.

Function Calling Intermediate

Constraining model outputs into a schema used to call external APIs/tools safely and deterministically.

Gating Network Intermediate

Chooses which experts process each token.

Generalization Intermediate

How well a model performs on new data drawn from the same (or similar) distribution as training.

Global Minimum Intermediate

Lowest possible loss.

Gradient Clipping Intermediate

Limiting gradient magnitude to prevent exploding gradients.

Gradient Descent Intermediate

Iterative method that updates parameters in the direction of negative gradient to minimize loss.

Gradient Leakage Intermediate

Recovering training data from gradients.

Gradient Noise Intermediate

Variability introduced by minibatch sampling during SGD.

Graph Attention Network Intermediate

GNN using attention to weight neighbor contributions dynamically.

Graph Convolution Intermediate

Extension of convolution to graph domains using adjacency structure.

Graph Neural Network Intermediate

Neural networks that operate on graph-structured data by propagating information along edges.

Grounding Intermediate

Constraining outputs to retrieved or provided sources, often with citation, to improve factual reliability.

Guardrails Intermediate

Rules and controls around generation (filters, validators, structured outputs) to reduce unsafe or invalid behavior.

Hallucination Intermediate

Model-generated content that is fluent but unsupported by evidence or incorrect; mitigated by grounding and verification.

Hessian Matrix Intermediate

Matrix of second derivatives describing local curvature of loss.

Heterogeneous Graph Intermediate

Graphs containing multiple node or edge types with different semantics.

Hidden Markov Model Intermediate

Probabilistic model for sequential data with latent states.

High-Frequency Trading Intermediate

Ultra-low-latency algorithmic trading.

High-Risk AI System Intermediate

AI used in sensitive domains requiring compliance.

Highway Network Intermediate

Early architecture using learned gates for skip connections.

Human Oversight Intermediate

Required human review for high-risk decisions.