Program Sampling

Program sampling aims to select representative program executions for efficient simulation or analysis, crucial for tasks like microprocessor design and AI-driven question answering. Current research focuses on improving sampling accuracy and efficiency, employing techniques like graph neural networks to learn program behavior from dynamic execution traces and integrating large language models to generate and evaluate informative questions within specific problem domains. These advancements promise faster hardware development cycles and more effective AI systems by reducing the computational cost and improving the quality of simulations and analyses.