M (54 terms)
Machine Learning
A subfield of AI where models learn patterns from data to make predictions or decisions, improving with experience ra...
Intermediate
MAP Estimation
Bayesian parameter estimation using the mode of the posterior distribution.
Intermediate
Marginalization
Eliminating variables by integrating over them.
Advanced
Market Design
Designing efficient marketplaces.
Advanced
Market Impact
Effect of trades on prices.
Advanced
Market Microstructure
Mechanics of price formation.
Intermediate
Markov Decision Process
Formal framework for sequential decision-making under uncertainty.
Intermediate
Masked Language Model
Predicts masked tokens in a sequence, enabling bidirectional context; often used for embeddings rather than generation.
Intermediate
Materials Discovery
AI discovering new compounds/materials.
Advanced
Maximum Likelihood Estimation
Estimating parameters by maximizing likelihood of observed data.
Intermediate
Mean Squared Error
Average of squared residuals; common regression objective.
Intermediate
Mechanism Design
Designing systems where rational agents behave as desired.
Advanced
Medical Imaging AI
AI applied to X-rays, CT, MRI, ultrasound, pathology slides.
Intermediate
Memory
Mechanisms for retaining context across turns/sessions: scratchpads, vector memories, structured stores.
Intermediate
Memory Augmentation
Extending agents with long-term memory stores.
Intermediate
Mental Simulation
Imagined future trajectories.
Frontier
Mesa-Optimizer
Learned subsystem that optimizes its own objective.
Advanced
Message Passing Neural Network
GNN framework where nodes iteratively exchange and aggregate messages from neighbors.
Intermediate
Meta-Cognition
Awareness and regulation of internal processes.
Frontier
Meta-Learning
Methods that learn training procedures or initializations so models can adapt quickly to new tasks with little data.
Intermediate
Mixed-Motive Game
Combination of cooperation and competition.
Advanced
Mixture of Experts
Routes inputs to subsets of parameters for scalable capacity.
Intermediate
MLOps
Practices for operationalizing ML: versioning, CI/CD, monitoring, retraining, and reliable production management.
Intermediate
Mode Collapse
Generator produces limited variety of outputs.
Advanced
Model
A parameterized mapping from inputs to outputs; includes architecture + learned parameters.
Intermediate
Model Card
Standardized documentation describing intended use, performance, limitations, data, and ethical considerations.
Intermediate
Model Disclosure
Requirement to reveal AI usage in legal decisions.
Intermediate
Model Documentation
Required descriptions of model behavior and limits.
Intermediate
Model Governance
Policies and practices for approving, monitoring, auditing, and documenting models in production.
Intermediate
Model Inventory
Central catalog of deployed and experimental models.
Intermediate
Model Inversion
Inferring sensitive features of training data.
Intermediate
Model Moat
Competitive advantage from proprietary models/data.
Intermediate
Model Orchestration
Coordinating models, tools, and logic.
Intermediate
Model Predictive Control
Optimizes future actions using a model of dynamics.
Intermediate
Model Registry
Central system to store model versions, metadata, approvals, and deployment state.
Intermediate
Model Release Control
Restricting distribution of powerful models.
Intermediate
Model Risk
Risk of incorrect financial models.
Intermediate
Model Risk Management
Framework for identifying, measuring, and mitigating model risks.
Intermediate
Model Stealing
Reconstructing a model or its capabilities via API queries or leaked artifacts.
Intermediate
Model Tiering
Classifying models by impact level.
Intermediate
Model Watermarking
Embedding signals to prove model ownership.
Intermediate
Model-Based RL
RL using learned or known environment models.
Advanced
Model-Free RL
RL without explicit dynamics model.
Advanced
Momentum
Uses an exponential moving average of gradients to speed convergence and reduce oscillation.
Intermediate
Monitoring
Observing model inputs/outputs, latency, cost, and quality over time to catch regressions and drift.
Intermediate
Monte Carlo Estimation
Approximating expectations via random sampling.
Advanced
Morphological Computation
Physical form contributes to computation.
Advanced
Motion Planning
Computing collision-free trajectories.
Advanced
Multi-Agent System
Multiple agents interacting cooperatively or competitively.
Intermediate
Multi-Head Attention
Allows model to attend to information from different subspaces simultaneously.
Intermediate
Multimodal Fusion
Combining signals from multiple modalities.
Intermediate
Multimodal Model
Models that process or generate multiple modalities, enabling vision-language tasks, speech, video understanding, etc.
Intermediate
Multitask Learning
Training one model on multiple tasks simultaneously to improve generalization through shared structure.
Intermediate
Mutual Information
Quantifies shared information between random variables.
Intermediate