AI-ordlista
Den kompletta ordlistan över AI
Gaussian Processes
Probabilistic models that define a distribution over functions, forming the basis of Bayesian optimization for modeling uncertainty.
Acquisition Functions
Heuristic strategies such as EI, UCB, and PI that balance exploration and exploitation to guide the selection of evaluation points.
Multi-Objective Optimization
Extension of BO to simultaneously optimize multiple conflicting objective functions by finding a Pareto front.
High-Dimensional Optimization
Specialized techniques such as ALEBO and ADD-GP to overcome the curse of dimensionality in BO.
Parallel and Batch Optimization
Methods that simultaneously evaluate multiple points to accelerate optimization using parallel resources.
Meta-Learning for BO
Approaches transferring knowledge from previous optimizations to accelerate new optimization tasks.
Neural Architecture Optimization
Application of BO to automatically discover optimal neural network architectures.
Hyperparameter Tuning
Using BO to automatically optimize machine learning model hyperparameters.
Multi-Fidelity Optimization
Strategies using low-cost approximations to accelerate the optimization of expensive functions.
BO with Constraints
Extensions of BO handling explicit or implicit constraints in the search space.
BO for Noisy Data
Techniques adapting BO to work with noisy or stochastic function evaluations.
BO for Contextual Bandits
Application of BO to contextual bandit problems where actions depend on context.
Kernel Design for BO
Design and selection of specialized kernels to effectively capture the structure of objective functions.
Bayesian Optimization for Combinatorial Optimization
Adaptations of Bayesian Optimization for discrete or combinatorial search spaces such as graphs and permutations.
Real-Time BO
BO implementations optimized for the strict time constraints of real-time applications.