Robots Learn to Correct Their Own Memory Errors: MIT Unveils UQ-DAAAM System
MIT scientists have developed the UQ-DAAAM system, which allows robots to identify and correct errors in their 3D memory independently. In trials, uncertainty was eliminated for 92% of objects at an average cost of two additional captions.
Read more →