MIT Researchers Introduce DAAAM for Robot Persistent Memory

MIT researchers introduced DAAAM last week to give robots persistent memory of environments they visit, forbes.com reported. The system builds detailed 4D maps, attaches descriptions to objects, and answers plain-English questions about earlier observations. DAAAM relies on a 4D scene graph that tracks objects, locations, and changes over time.

In tests the approach raised accuracy between 21 and 53 percent compared with earlier methods, depending on the question and task. The work was presented at the Conference on Computer Vision and Pattern Recognition. Lead author Nicolas Gorlo, an MIT graduate student, worked with co-author Lukas Schmid, now a professor at the University of Technology Nuremberg in Germany, and Luca Carlone, an associate professor in MIT’s Department of Aeronautics and Astronautics who directs the MIT SPARK Laboratory.

Carlone said robots must reason about time and space the way humans do if they are to work alongside people. ” A follow-on paper, UQ-DAAAM, adds uncertainty quantification so the system can flag potentially unreliable object descriptions. Professional service robot sales reached nearly 200,000 units in 2024, up 9 percent from the prior year, with transportation and logistics accounting for 102,900 units, according to the International Federation of Robotics.

Those figures illustrate environments where persistent memory could reduce errors and human interventions.