Search papers, labs, and topics across Lattice.
This paper introduces the Sequential RC-TGAN (Seq. RC-TGAN), an advanced framework for generating synthetic relational time series that effectively captures complex temporal dynamics, particularly in categorical sequences. By employing a novel integrated loss function based on Spectral Envelope Theory, the model optimizes the preservation of latent periodic structures during training, addressing the limitations of standard encoding methods. Experimental results reveal that Seq. RC-TGAN significantly outperforms existing systems in accurately reproducing cyclic patterns and long-term seasonality in both categorical and continuous time series data.
Spectral Envelope Theory unlocks a new level of fidelity in generating synthetic relational time series, outperforming traditional methods in capturing complex temporal dynamics.
The generation of synthetic relational databases often involves modeling complex temporal dynamics, such as transaction logs or event sequences. A significant challenge in this domain is the handling of categorical time series (e.g., status codes), where standard encoding methods like one-hot encoding fail to capture intrinsic frequency-domain features such as seasonality and cyclicity. In this paper, we introduce Sequential RC-TGAN (Seq. RC-TGAN), a temporal extension of the RC-TGAN framework, equipped with a novel integrated loss function based on the \textit{Spectral Envelope Theory}. This differentiable loss allows the generator to directly optimize the preservation of latent periodic structures via backpropagation. While spectral envelope theory is inherently designed for categorical sequences, we extend this frequency-domain regularization to continuous time series by employing a Variational Gaussian Mixture Model (VGM) discretization strategy. To establish a mathematically rigorous evaluation standard, we simulate categorical time series governed by a parameter $伪$, with exactly known theoretical spectral envelopes. Integrating these dynamic sequences into the child tables of a relational database yields a robust ground-truth benchmark for evaluating the frequency-domain fidelity of our generative framework. Furthermore, we address the lack of robust evaluation standards for relational time series by proposing two new metrics: Spectral Density Divergence and Spectral Envelope Divergence. Experimental results on real-world datasets, as well as our simulated benchmarks, demonstrate that our end-to-end approach significantly outperforms state-of-the-art systems in reproducing cyclic patterns and long-term seasonality across both categorical and continuous features.
Mila