Check out research publications produced under the PlanetAI Research Lab focusing on sustainable artificial intelligence, responsible computing, Human-AI Cointelligence and environmentally aware digital infrastructure.
This paper proposes the usage of carbon-aware training schedules as an efficient Green AI technique that helps to significantly decrease the energy demand as well as CO2 emissions associated with machine learning models without impairing their prediction performance.
View PublicationAcknowledging energy as a computational resource, the study introduces Energy Complexity as a hardware-parameterized computational measure which extends the concept of classical operation counting by weighting arithmetic, memory, and communication events according to their physical energy cost.
Coming SoonThis paper proposes Green Activation Functions—energy-aware nonlinearities designed to reduce the computational energy demand of deep learning models. We examine how seemingly small architectural choices can influence the overall environmental footprint of DL.
Coming SoonThis case study examines the ethical and societal risks arising from bias in facial recognition systems, where differences in training data can lead to unequal accuracy across demographic groups. It highlights the importance of fairness auditing, transparent datasets, and responsible AI design to mitigate discrimination and improve trust in automated decision-making systems.
Study CaseAI sustainability discussions often focus on making data centers more energy-efficient, but the environmental impact of AI also depends on model design, training strategies, and algorithmic complexity. True sustainable AI therefore requires optimizing the entire AI lifecycle—from data and algorithms to deployment and usage—not just the infrastructure that runs it.
Study Opinion