What question did this study set out to answer?

The aim is to create a framework that minimizes energy consumption and carbon emissions in AI models while maintaining accuracy.

February 26, 2026Open Access

CAS-NAS: A carbon-aware neural architecture search framework for sustainable AI development

Q: What does this research mean for the field?

The CAS-NAS framework achieves 30%–42% reductions in energy consumption and 28%–38% lower carbon emissions compared to ResNet-50 and EfficientNet baselines while maintaining competitive accuracy. Novelty: ClaimNovelty.NOVEL_FINDING. Consensus alignment: ConsensusAlignment.ESTABLISHES_NEW_DIRECTION.

Puntos clave

The aim is to create a framework that minimizes energy consumption and carbon emissions in AI models while maintaining accuracy.
Developed a carbon-aware neural architecture search framework using NSGA-II-based optimization.
Utilized real-time carbon intensity simulation and hardware-aware energy estimation.
Evaluated models using proxy datasets (CIFAR-10/100) and various hardware profiles.
Achieved 30%–42% reductions in energy consumption and 28%–38% reduction in carbon emissions compared to baseline models.
Demonstrated a trade-off where a 1% accuracy reduction led to 45%–60% energy savings.
Established standardized metrics for Green AI, enhancing sustainability comparisons.

Resumen

The exponential growth of artificial intelligence has precipitated significant environmental concerns, with large-scale models consuming massive computational resources and generating substantial carbon emissions. This research addresses this critical challenge through a novel carbon-aware neural architecture search (CAS-NAS) framework that optimizes deep learning models across three objectives: predictive accuracy, energy efficiency, and carbon footprint. Our NSGA-II-based optimization demonstrates 30%–42% reductions in energy consumption (Joules/inference) and 28%–38% lower carbon emissions (gCO 2 e) compared to ResNet-50 and EfficientNet baselines, while maintaining competitive accuracy (less than 2% drop). The methodology integrates real-time carbon intensity simulation with hardware-aware energy estimation and multi-objective evolutionary algorithms to discover Pareto-optimal architectures. Our comprehensive evaluation system uses proxy datasets (CIFAR-10/100) and hardware profiles (Raspberry Pi 4, NVIDIA Jetson) to simulate real-world conditions without physical data collection. The framework incorporates sparsity-driven quantization, dynamic precision scaling, and carbon-aware scheduling to adapt models to variable grid conditions. Experimental results reveal critical trade-offs: a 1% accuracy reduction yields 45%–60% energy savings, while temporal carbon shifting decreases emissions by 40% during renewable energy availability. Our work establishes the first standardized evaluation metrics for Green Artificial Intelligence (Green AI): Carbon Efficiency Score (accuracy per gCO 2 e) and Energy-Accuracy Ratio enabling direct comparison of sustainable algorithms. This research provides both a methodological blueprint for environmentally conscious AI development and empirical evidence that substantial emission reductions are achievable through algorithmic optimization, supporting global climate initiatives like the EU Green Deal. The implementation is publicly available to accelerate adoption in edge computing and cloud infrastructure.

Leer artículo completoexternamente

Me gusta

Guardar

Ver artículo completo