Sustainable AI and User-Friendly Data Visualization

The research landscape in the field of energy efficiency and sustainability in AI and IoT is rapidly evolving, with a strong emphasis on developing practical tools and methodologies that address the growing concerns about energy consumption and data complexity. Recent studies highlight the critical need for more sustainable AI practices, particularly in the context of deep learning models, where the pursuit of marginal accuracy gains often comes at a steep energy cost. Innovative solutions are being proposed to improve energy efficiency, including the introduction of scoring systems and interactive applications that facilitate informed decision-making. Additionally, there is a growing focus on creating user-friendly, immersive data visualization tools that cater to a broad audience, enabling better understanding and actionable insights from complex datasets. These developments underscore a shift towards more sustainable and efficient AI technologies, as well as the democratization of data visualization tools to support real-world applications in energy management and beyond.

Sources

Double-Exponential Increases in Inference Energy: The Cost of the Race for Accuracy

Illustrating Transition Scenarios to Renewable Energy in Hawaii with ProjecTable

BiTSA: Leveraging Time Series Foundation Model for Building Energy Analytics

Built with on top of