Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a systematic framework for tabular data disentanglement, addressing the challenges of complex attribute interactions and limitations of existing methods like factor analysis and VAEs. The framework modularizes the disentanglement process into data extraction, data modeling, model analysis, and latent representation extrapolation. A case study on synthetic tabular data generation demonstrates the framework's applicability and potential for data synthesis.
Tabular data disentanglement, often an afterthought, gets a systematic framework that could unlock more effective and efficient processing across domains like finance and industrial control.
Tabular data, widely used in various applications such as industrial control systems, finance, and supply chain, often contains complex interrelationships among its attributes. Data disentanglement seeks to transform such data into latent variables with reduced interdependencies, facilitating more effective and efficient processing. Despite the extensive studies on data disentanglement over image, text, or audio data, tabular data disentanglement may require further investigation due to the more intricate attribute interactions typically found in tabular data. Moreover, due to the highly complex interrelationships, direct translation from other data domains results in suboptimal data disentanglement. Existing tabular data disentanglement methods, such as factor analysis, CT-GAN, and VAE face limitations including scalability issues, mode collapse, and poor extrapolation. In this paper, we propose the use of a framework to provide a systematic view on tabular data disentanglement that modularizes the process into four core components: data extraction, data modeling, model analysis, and latent representation extrapolation. We believe this work provides a deeper understanding of tabular data disentanglement and existing methods, and lays the foundation for potential future research in developing robust, efficient, and scalable data disentanglement techniques. Finally, we demonstrate the framework's applicability through a case study on synthetic tabular data generation, showcasing its potential in the particular downstream task of data synthesis.