Symbolic-KANs bridge the gap between scalable deep learning and interpretable symbolic regression by embedding discrete library primitives directly into the network.
March 26, 2026
Original Paper
Symbolic--KAN: Kolmogorov-Arnold Networks with Discrete Symbolic Structure for Interpretable Learning
arXiv · 2603.23854
The Takeaway
This architecture allows for the discovery of exact governing equations within a trainable deep network. It avoids the combinatorial search of classical symbolic regression while maintaining the efficiency of Kolmogorov-Arnold Networks, enabling the recovery of closed-form expressions for complex dynamical systems.
From the abstract
Symbolic discovery of governing equations is a long-standing goal in scientific machine learning, yet a fundamental trade-off persists between interpretability and scalable learning. Classical symbolic regression methods yield explicit analytic expressions but rely on combinatorial search, whereas neural networks scale efficiently with data and dimensionality but produce opaque representations. In this work, we introduce Symbolic Kolmogorov-Arnold Networks (Symbolic-KANs), a neural architecture