Physics-Based Engineering and the Machine-Learning "Black Box" Problem


In MIT 2.C161, George Barbastathis demonstrates how mechanical engineers can use their knowledge of physical systems to keep algorithms in check and develop more accurate predictions. Machine-learning algorithms are often referred to as a "black box." Once data are put into an algorithm, it's not always known exactly how the algorithm arrives at its prediction. This can be particularly frustrating when things go wrong. A new mechanical engineering (MechE) course at MIT teaches students how to tackle the "black box" problem, through a combination of data science and physics-based engineering.

Social distancing surveillance: how robots will keep you in line


WiFi and Bluetooth-based methods are accurate in detecting social distancing breaches and our approach complements them. The WiFI and Bluetooth-based methods need appropriate sensing technologies and cannot be easily deployed in all kinds of environments (e.g., public places or isolated locations). These methods also need additional infrastructure to be in place for detection. Our method uses the visual feed from a depth camera onboard a mobile robot and existing CCTV infrastructure (if available) to detect social distancing breaches. In addition, the robot can autonomously navigate and interact with people and encourage them to maintain social distancing.