Glossary

Markov Decision Process (MDP) refers to a statistical theory that models scenarios where outcomes are partly based on random events and partly on calculated decisions. This method extends from Markov chains with added rewards and actions, enabling choices and inducing motivation.

Mathematical Optimization Methods involve mathematical programming to find the optimal solution among alternatives, under discrete or continuous optimization. This involves maximizing or minimizing scalar functions.

Improve your LLM Stack Layers with our series of Performance Optimization methods. From pruning to distillation, prompt engineering, and specialized libraries, optimize your model efficiency.