Cambridge Yang Graduates

Cambridge Yang recently graduated from the Massachusetts Institute of Technology with his Ph.D. in Electrical Engineering and Computer Science, completing his dissertation on “On the Learnability of General Reinforcement-learning Objectives” under the supervision of Professor Michael Carbin. Yang’s research addresses fundamental limitations in reinforcement learning by moving beyond traditional reward-based approaches to establish a theoretical framework for specifying and learning general objectives.  Yang’s contributions help advance the theoretical foundations for designing intelligent agents that can align with formally defined objectives beyond the limitations of reward-based surrogates. His research has significant implications for AI safety and alignment, particularly in high-stakes applications where formal guarantees about system behavior are critical.