The paper introduces SigKAN, a novel approach that enhances multivariate function approximation using learnable path signatures and Kolmogorov-Arnold networks (KANs). By weighting the values obtained by KANs using learnable path signatures, which capture important geometric features of paths, SigKAN improves the learning capabilities of these networks. This combination allows for a more comprehensive and flexible representation of sequential and temporal data. The authors demonstrate that SigKAN outperforms conventional methods in various function approximation challenges, particularly in time series analysis and forecasting. The method leverages path signatures in neural networks to enhance performance, offering intriguing opportunities for applications in fields such as financial modeling and time series analysis. The paper also provides detailed architectural descriptions, including the Gated Residual KAN (GRKAN) and Learnable Path Signature layers, and presents experimental results showing the effectiveness of SigKAN in predicting volumes and absolute returns on the Binance exchange.The paper introduces SigKAN, a novel approach that enhances multivariate function approximation using learnable path signatures and Kolmogorov-Arnold networks (KANs). By weighting the values obtained by KANs using learnable path signatures, which capture important geometric features of paths, SigKAN improves the learning capabilities of these networks. This combination allows for a more comprehensive and flexible representation of sequential and temporal data. The authors demonstrate that SigKAN outperforms conventional methods in various function approximation challenges, particularly in time series analysis and forecasting. The method leverages path signatures in neural networks to enhance performance, offering intriguing opportunities for applications in fields such as financial modeling and time series analysis. The paper also provides detailed architectural descriptions, including the Gated Residual KAN (GRKAN) and Learnable Path Signature layers, and presents experimental results showing the effectiveness of SigKAN in predicting volumes and absolute returns on the Binance exchange.