Our mission is to advance artificial intelligence by understanding and engineering how knowledge is learned, represented, retained, and used within AI models.

Rather than focusing solely on building ever-larger models, we aim to develop AI that can accumulate knowledge over time, adapt to changing environments, and make effective use of what it has learned. Toward this goal, we pursue AI models that are:

• (Accessible) Learned knowledge can be understood, extracted, transferred, and reused. 

• (Scalable) Models can efficiently learn, retain, and adapt knowledge throughout their lifetime.  

• (Reliable) Learned knowledge can be examined, corrected, reorganized, or removed when necessary. 

From an academic perspective, we pursue research that is theoretically rigorous, computationally efficient, and practically relevant. We aim to produce original and impactful contributions and publish our work in top-tier international venues in artificial intelligence and big data. Please see our Publications for our research outcomes.