Dianhuan Lin
Papers
2
Total Citations
258
H-Index
2
About
Dianhuan Lin is a leading researcher at the intersection of machine learning, inductive logic programming, and robotics. Her work centers on enabling machines to learn abstract concepts and transfer knowledge across vastly different domains—a crucial step toward more intelligent, adaptable AI. Lin’s foundational contribution, *Meta-interpretive learning of higher-order dyadic datalog: predicate invention revisited* (192 citations), revolutionized the field by showing how machines can invent new predicates from data, effectively learning to learn. This breakthrough in predicate invention has become a cornerstone for modern symbolic AI. More recently, Lin’s research on *Beyond imitation: Zero-shot task transfer on robots by learning concepts as cognitive programs* (66 citations) demonstrated a remarkable capability: robots that can infer high-level concepts from simple image pairs and apply them in entirely new physical contexts—such as assembling IKEA furniture from a diagram without prior training. This work bridges the gap between human-like conceptual reasoning and robotic execution, offering a path toward robots that truly understand our intent. Lin’s achievements have earned her recognition as a pioneer in cognitive robotics and program induction.
Research Focus
Key Achievements
Top Papers
- 1
- 2