Timezone: »

Learning Enhanced Representation for Tabular Data via Neighborhood Propagation
Kounianhua Du · Weinan Zhang · Ruiwen Zhou · Yangkun Wang · Xilong Zhao · Jiarui Jin · Quan Gan · Zheng Zhang · David P Wipf


Prediction over tabular data is an essential and fundamental problem in many important downstream tasks. However, existing methods either take a data instance of the table independently as input or do not fully utilize the multi-row features and labels to directly change and enhance the target data representations. In this paper, we propose to 1) construct a hypergraph from relevant data instance retrieval to model the cross-row and cross-column patterns of those instances, and 2) perform message Propagation to Enhance the target data instance representation for Tabular prediction tasks. Specifically, our specially-designed message propagation step benefits from 1) the fusion of label and features during propagation, and 2) locality-aware multiplicative high-order interaction between features. Experiments on two important tabular prediction tasks validate the superiority of the proposed PET model against other baselines. Additionally, we demonstrate the effectiveness of the model components and the feature enhancement ability of PET via various ablation studies and visualizations. The code is available at https://github.com/KounianhuaDu/PET.

Author Information

Kounianhua Du (Shanghai Jiao Tong University)
Weinan Zhang (Shanghai Jiao Tong University)
Ruiwen Zhou (Shanghai Jiao Tong University)
Yangkun Wang (Shanghai Jiao Tong University)
Xilong Zhao (Shanghai Jiaotong University)
Jiarui Jin (Shanghai Jiao Tong University)
Quan Gan (New York University)
Zheng Zhang (Shanghai New York Univeristy)
David P Wipf (AWS)

More from the Same Authors