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

arXiv: https://arxiv.org/abs/2602.02056 · HTML: https://mlsanity.com/paper/2602.02056/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 39 references that are in the map.
- First posted on arXiv 2 Feb 2026.

## Builds on (most cited first, 24 of 24)

- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [The Hardware Lottery](https://mlsanity.com/paper/2009.06489/) (2020)
- [All you need is a good init](https://mlsanity.com/paper/1511.06422/) (2015)
- [Deep Online Learning via Meta-Learning: Continual Adaptation for Model-Based RL](https://mlsanity.com/paper/1812.07671/) (2018)
- [hls4ml: An Open-Source Codesign Workflow to Empower Scientific Low-Power Machine Learning Devices](https://mlsanity.com/paper/2103.05579/) (2021)
- [Kolmogorov-Arnold Networks: A Critical Assessment of Claims, Performance, and Practical Viability](https://mlsanity.com/paper/2407.11075/) (2024)
- [Exploring the Limitations of Kolmogorov-Arnold Networks in Classification: Insights to Software Training and Hardware Implementation](https://mlsanity.com/paper/2407.17790/) (2024)
- [KANEL'E: Kolmogorov-Arnold Networks for Efficient LUT-based Evaluation](https://mlsanity.com/paper/2512.12850/) (2025)
- [Hardware Acceleration of Kolmogorov-Arnold Network (KAN) for Lightweight Edge Inference](https://mlsanity.com/paper/2409.11418/) (2024)
- [On-FPGA Training with Ultra Memory Reduction: A Low-Precision Tensor Method](https://mlsanity.com/paper/2104.03420/) (2021)
- [EF-Train: Enable Efficient On-device CNN Training on FPGA Through Data Reshaping for Online Adaptation or Personalization](https://mlsanity.com/paper/2202.10935/) (2022)
- [End-to-end workflow for machine learning-based qubit readout with QICK and hls4ml](https://mlsanity.com/paper/2501.14663/) (2025)
- [Automatic Compiler Based FPGA Accelerator for CNN Training](https://mlsanity.com/paper/1908.06724/) (2019)
- [Low latency optical-based mode tracking with machine learning deployed on FPGAs on a tokamak](https://mlsanity.com/paper/2312.00128/) (2023)
- [A survey on FPGA-based accelerator for ML models](https://mlsanity.com/paper/2412.15666/) (2024)
- [Learning Sparse Representations in Reinforcement Learning with Sparse Coding](https://mlsanity.com/paper/1707.08316/) (2017)
- [A Highly Parallel FPGA Implementation of Sparse Neural Network Training](https://mlsanity.com/paper/1806.01087/) (2018)
- [FPGA Implementation of Multi-Layer Machine Learning Equalizer with On-Chip Training](https://mlsanity.com/paper/2212.03515/) (2022)
- [FPGA-based Distributed Union-Find Decoder for Surface Codes](https://mlsanity.com/paper/2406.08491/) (2024)
- [Efficient Implementation of a Multi-Layer Gradient-Free Online-Trainable Spiking Neural Network on FPGA](https://mlsanity.com/paper/2305.19468/) (2023)
- [Demonstrating the Potential of Adaptive LMS Filtering on FPGA-Based Qubit Control Platforms for Improved Qubit Readout in 2D and 3D Quantum Processing Units](https://mlsanity.com/paper/2408.00904/) (2024)
- [Function Approximation Using Analog Building Blocks in Flexible Electronics](https://mlsanity.com/paper/2502.01489/) (2025)
- [Online Learning Extreme Learning Machine with Low-Complexity Predictive Plasticity Rule and FPGA Implementation](https://mlsanity.com/paper/2512.21777/) (2025)

## Abstract (arXiv)

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 Multi-Layer Perceptrons (MLPs) are both inefficient and numerically unstable. We identify 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 than MLPs across a range of low-latency and resource-constrained tasks. To our knowledge, this work is the first to demonstrate model-free online learning at sub-microsecond latencies.