ECOFRIENDLY ARTIFICIAL INTELLIGENCE SYSTEM FOR LOW POWER DEVICES - REVIEW ARTICLE

As AI technologies advance in low-power devices, developing eco-friendly and energy-efficient solutions is crucial. This paper explores methods to create AI systems that reduce environmental impact while maintaining high performance. Key approaches include model pruning, which simplifies networks by removing unnecessary parameters; quantization, which reduces memory and processing demands through lower-precision computations; and knowledge distillation, which transfers capabilities from larger models to more compact ones. Case studies demonstrate these methods\' effectiveness in enhancing efficiency and reducing power consumption across applications like mobile devices and edge computing. These strategies help balance performance with environmental considerations, making AI systems better suited for low-power environments.