A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates a novel method for training AI models to perform complex tasks like robot navigation and manipulation. The approach, called ‘Consensus Game,’ treats the training process as a game between two AI components: one that generates answers and another that evaluates them. This adversarial setup helps the model learn more robustly from unlabeled data, reducing the need for vast, manually curated datasets. The research shows promising results in improving the efficiency and accuracy of AI systems in interpreting and acting on visual and textual instructions. For the full details, read the complete article at https://technologyreview.com/2024/07/18/1094815/mit-ai-consensus-game-training/.