A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates a novel approach to training AI models that significantly reduces energy consumption. The method, called ‘Liquid Neural Networks,’ uses a more biologically inspired architecture that activates only relevant parts of the network for a given task, unlike traditional models that run all computations regardless of necessity. Early tests show this technique can cut energy use during training by up to 80% while maintaining, and in some cases improving, model accuracy on tasks like image classification and robotic control. Researchers highlight the potential for this advancement to make AI development more sustainable and accessible, especially for resource-constrained environments. The full details of the research are available in the published paper. Read the full article at https://technologyreview.com/2024/03/15/liquid-networks-energy-efficient-ai.