Fast Machine Learning for Science Conference 2026

Ultrafast On-chip Online Learning via Spline Locality in Kolmogorov-Arnold Networks

Not scheduled

12m

Presentation Contributed Talks

Speaker

Philip Coleman Harris (Massachusetts Inst. of Technology (US))

Description

Ultrafast online learning is essential for high-frequency systems, such as controls for quantum computing and nuclear fusion, where adaptation must occur on sub-microsecond timescales. Meeting these requirements demands low-latency, fixed-precision computation under strict memory constraints, a regime in which conventional networks buil on Multi-Layer Perceptrons (MLPs) are both inefficient and numerically unstable. We present a pathway towards stable low-precision training on FPGAs, and demonstrate its performance in the online learning scenario. Our pathway is built on key properties of Kolmogorov-Arnold Networks (KANs) that align with these constraints. Specifically, we show that: (i) KAN updates exploiting B-spline locality are sparse, enabling superior on-chip resource scaling, and (ii) KANs are inherently robust to fixed-point quantization. By implementing fixed-point online training on Field-Programmable Gate Arrays (FPGAs), a representative platform for on-chip computation, we demonstrate that KAN-based online learners are significantly more efficient and expressive for real-time learning. Our approach opens the door to many new possible scientific applications, and we showcase this in the context of real-time controls of quantum computers. To our knowledge, this work is the first to demonstrate model-free online learning at sub-microsecond latencies.