Much of Youcheng’s recent research work is about testing and verification of deep neural networks (DNNs). Overall, He is interested in building trustworthy software. Youcheng is among the first to study the testing of deep neural networks (DNNs). He proposed MC/DC-style test criteria for testing DNNs, and developed the concolic testing approach for test case generation of neural networks. He is the first researcher to use software fault localisation techniques for explaining DNNs.