Kolmogorov-Arnold Networks
PulseAugur coverage of Kolmogorov-Arnold Networks — every cluster mentioning Kolmogorov-Arnold Networks across labs, papers, and developer communities, ranked by signal.
6 天有情绪数据
-
New KAN variants tackle efficiency and hardware implementation
Researchers have developed a new variant of Kolmogorov-Arnold Networks (KANs) called Kolmogorov-Arnold Fourier Networks (KAFs) to address limitations in parameter efficiency and high-frequency feature capture. KAFs repa…
-
K-U-KAN reconstructs 3D dental models from single X-rays
Researchers have developed K-U-KAN, a novel three-stage pipeline for reconstructing 3D dental models from single panoramic X-ray images. This method utilizes Kolmogorov-Arnold Networks (U-KAN) enhanced with Koopman oper…
-
Holomorphic KAN-ODE 模型以可解释方程模拟复杂动力学
研究人员开发了一个名为 Holomorphic KAN-ODE 的新框架，将 Kolmogorov-Arnold Networks (KANs) 集成到神经常微分方程 (Neural ODEs) 中。该方法通过纳入复分析先验并遵守 Cauchy-Riemann 条件，旨在更好地模拟具有分形边界的复杂动力学系统。与传统的 MLP 相比，Holomorphic KAN-ODE 框架表现出卓越的性能，在重建动力学系统、识别控制方程以及提高对…
-
KANs 推动生存分析，提出新型深度学习模型
研究人员开发了 KAPLAN-HR，一种基于 Kolmogorov-Arnold Networks (KANs) 的新型深度学习模型，用于生存分析。该模型可以估计协变量和时间的联合函数作为条件风险率，克服了传统方法需要手动指定复杂效应的局限性。在六个临床数据集上的评估表明，KAPLAN-HR 的性能与现有的统计和深度学习生存分析技术相当或更优。
-
混合KAN-XGBoost模型改进电力价格预测
研究人员开发了一种新的混合框架，用于预测澳大利亚国家电力市场（NEM）的电力价格。该方法结合了Kolmogorov-Arnold网络（KAN）和XGBoost，以更好地捕捉复杂的市场动态，包括由高可再生能源渗透率加剧的波动性和价格飙升。实验表明，与LSTM和独立的KAN或XGBoost等现有方法相比，该混合模型表现显著更优，与单独使用XGBoost相比，平均绝对误差（MAE）降低了约12%。
-
新算法优化随机神经网络中的激活函数
研究人员开发了一种新算法，用于优化随机神经网络（RaNNDy）中的激活函数，以逼近动力系统中的传递算子。该方法保持网络权重和偏置固定，显著降低了训练成本，同时提高了基函数的适用性。另外，一篇综述回顾了神经网络逼近理论的演变，涵盖了经典的密度结果、逼近误差的定量界限，以及深度和宽度等架构特征的影响。它还强调了近期对Kolmogorov-Arnold Networks (KANs) 作为一种替代架构范式的关注。
-
Adaptive RBF-KAN improves efficiency with new kernels and data-driven shape parameters
Researchers have developed an enhanced version of Kolmogorov-Arnold Networks (KANs) called adaptive RBF-KAN, which improves computational efficiency and flexibility. This new approach replaces the fixed Gaussian radial …
-
New theory bounds KAN training, reveals privacy-utility gap
Researchers have established new theoretical bounds for training Kolmogorov-Arnold Networks (KANs), a structured alternative to standard MLPs. The work analyzes KANs trained with mini-batch stochastic gradient descent (…
-
KAN-CL framework reduces catastrophic forgetting in continual learning
Researchers have introduced KAN-CL, a new framework for continual learning that addresses catastrophic forgetting by leveraging the unique structure of Kolmogorov-Arnold Networks (KANs). This method applies importance-w…
-
Bayesian KANs achieve near-minimax rates in new theory
Researchers have developed a theoretical framework for sparse Bayesian Kolmogorov-Arnold Networks (KANs). Their work establishes statistical foundations for KANs, demonstrating that these networks can achieve near-minim…
-
Kolmogorov-Arnold Networks evolve with automated basis learning and practitioner's guide
Researchers have introduced InfinityKAN, a novel framework that automates the selection of basis functions in Kolmogorov-Arnold Networks (KANs), a theoretically grounded alternative to traditional multi-layer perceptron…
-
Neural Operators advance interpolation, resolution robustness, and Bayesian inference
Researchers are exploring new applications and improvements for neural operators, a class of models designed for learning maps between function spaces. One paper reframes neural operators as efficient function interpola…
-
KANs gain temporal explanations with new Temporal Functional Circuits
Researchers have developed a new framework called Temporal Functional Circuits to enhance the interpretability of Kolmogorov-Arnold Networks (KANs) in time-series forecasting. This method transforms the KAN's internal e…
-
量子启发式特征求解器大幅减少参数，提升量子化学性能
研究人员开发了一种名为GQKAE的新型量子启发式特征求解器，旨在提高量子化学领域高性能计算的效率。该模型用混合量子启发式Kolmogorov-Arnold网络模块取代了传统的馈通网络，可将可训练参数和内存使用量显著减少约66%。基准测试表明，GQKAE在实现与现有GPT基方法相当的化学精度方面，同时为复杂系统提供了更优的收敛性和能量误差。
-
KANs enable ultrafast on-chip online learning for low-latency systems
Researchers have demonstrated ultrafast online learning capabilities using Kolmogorov-Arnold Networks (KANs) on Field-Programmable Gate Arrays (FPGAs). This approach achieves sub-microsecond adaptation times, outperform…
-
P1-KAN network offers improved accuracy and convergence over MLPs
Researchers have introduced P1-KAN, a novel Kolmogorov-Arnold Network designed to approximate complex, irregular functions in high-dimensional spaces. The paper provides theoretical error bounds and universal approximat…
-
SRGAN-CKAN improves image super-resolution with efficient local operators
Researchers have developed SRGAN-CKAN, a novel framework for single-image super-resolution that enhances local operators for improved detail reconstruction. This approach integrates Convolutional Kolmogorov-Arnold Netwo…
-
新AI方法提升时间序列预测的准确性和可解释性
研究人员引入了几种新的时间序列预测方法，旨在提高准确性和泛化能力。MeLISA是一种无潜在变量的自回归模型，可提高回溯效率和长视界统计准确性。Temporal Functional Circuits利用Kolmogorov-Arnold Networks (KANs)为预测提供忠实且与时间相关的解释。Dynamic Pattern Recalibration (DPR)提供了一种与骨干网络无关的令牌级重新校准机制，以适应不断变化的局部…
-
New penalty method enhances KAN interpretability without sacrificing accuracy
Researchers have developed a new curvature penalty for Kolmogorov-Arnold Networks (KANs) to address issues with high-curvature oscillations in their activation functions. This penalty aims to improve the interpretabilit…
-
New research details Lipschitz-product control for deep KAN representations
Researchers have developed a method for deep Kolmogorov-Arnold Network (KAN) representations of complex functions, ensuring a layer-wise Lipschitz product control. This approach guarantees a domain-sensitive bound indep…