Professor Lyu Chen and his team at Nanyang Technological University (NTU) have developed a self-driving vehicle capable of continual learning. After being deployed, the vehicle can learn from human drivers' interventions and corrections when they take over. The related research, titled "Human-guided continual learning for multifaceted improvement of self-driving vehicles," was recently published in the journal Nature Communications. In operation, the vehicle handles the driving while humans supervise and intervene to correct mistakes. The team designed the system to both absorb new corrections and preserve existing skills; when the vehicle's judgment diverges from human operation, the system weighs the correction against existing skills before adopting it. In real-track field tests, after several rounds of human-guided corrections, the vehicle's "ethical risk" metric dropped to about one-tenth of its initial level. After learning from corrections in a few rare scenarios, the vehicle's performance also improved in other rare scenarios it had not encountered before. Currently, the system is designed to drive by following rules and does not yet have the ability to break them. The research team notes that how to accurately distinguish "valid corrections" from "erroneous interventions" by human drivers remains an unresolved issue.