Planning Based Compilation

Planning-based compilation focuses on optimizing the translation of high-level descriptions of computations (like deep neural networks or quantum algorithms) into efficient low-level implementations for specific hardware platforms. Current research emphasizes optimizing for resource-constrained environments (e.g., edge devices, embedded systems), exploring techniques like neural architecture search, tensor optimization, and reinforcement learning to achieve improved performance and reduced resource consumption for various models, including vision transformers and convolutional neural networks. This work is crucial for deploying computationally intensive machine learning and quantum computing applications across a wider range of devices and significantly impacts the efficiency and scalability of these technologies.