October 2026 — Prof. Cyrus Shahabi was recently featured in USC Viterbi School of Engineering News, discussing how artificial intelligence can help cities adapt to rapidly changing road networks following major earthquakes.

The article, titled “After an Earthquake, the Road Map Changes. This AI Model Was Built for That,” highlights the research team’s development of DeepStateGNN, a novel graph neural network model designed to improve traffic prediction under changing road network conditions. By leveraging semantic representations rather than relying on fixed road network structures, DeepStateGNN demonstrates strong generalization to previously unseen sensors while achieving improved computational efficiency.

The research highlights the potential of AI-driven traffic modeling to support real-time emergency response, enabling faster evaluation of alternative traffic management strategies during disasters.

Read the full article: USC Viterbi News – After an Earthquake, the Road Map Changes. This AI Model Was Built for That