A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has developed a machine learning framework that can predict the future ‘surprise’ of an image or video sequence. The system, trained on hundreds of thousands of videos, learns a model of what makes a visual event predictable or surprising based on contextual clues. In tests, the model successfully identified unusual occurrences, such as a car driving on grass instead of a road, with greater accuracy than previous methods. This research aims to improve AI’s common sense understanding of the physical world, which is crucial for applications like autonomous vehicles and robotic assistants. The work was published in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Read the full article at https://technologyreview.com/2023/06/15/1074568/ai-predict-surprise-videos/