A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates a significant advancement in robotic manipulation. Researchers have developed a system that allows robots to perform complex, dexterous tasks, such as hanging mugs on a rack or placing bowls on a shelf, by using a more human-like approach to planning and executing movements. The system, named PIGINet (Plans, Images, Goals, and Initial Facts), employs a transformer-based neural network to evaluate and refine task plans. It focuses on identifying and eliminating infeasible plans early in the process, which drastically reduces planning time. In tests within simulated home environments, the approach cut down planning time by 50% and increased the success rate of task completion by 40-50%. This work represents a step toward more efficient and capable household robots that can better navigate cluttered and unpredictable spaces. For the full details, read the article at https://technologyreview.com/2024/07/12/1094755/robots-plan-tasks-like-we-do/