English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Sequential tabular data is one of the most commonly used data types inreal-world applications. Different from conventional tabular data, where rowsin a table are independent, sequential tabular data contains rich contextualand sequential information, where some fields are dynamically changing overtime and others are static. Existing transformer-based approaches analyzingsequential tabular data overlook the differences between dynamic and staticfields by replicating and filling static fields into each transformer, andignore temporal information between rows, which leads to three majordisadvantages: (1) computational overhead, (2) artificially simplified data formasked language modeling pre-training task that may yield less meaningfulrepresentations, and (3) disregarding the temporal behavioral patterns impliedby time intervals. In this work, we propose FATA-Trans, a model with two fieldtransformers for modeling sequential tabular data, where each processes staticand dynamic field information separately. FATA-Trans is field- and time-awarefor sequential tabular data. The field-type embedding in the method enablesFATA-Trans to capture differences between static and dynamic fields. Thetime-aware position embedding exploits both order and time interval informationbetween rows, which helps the model detect underlying temporal behavior in asequence. Our experiments on three benchmark datasets demonstrate that thelearned representations from FATA-Trans consistently outperformstate-of-the-art solutions in the downstream tasks. We also presentvisualization studies to highlight the insights captured by the learnedrepresentations, enhancing our understanding of the underlying data. Our codesare available at https://github.com/zdy93/FATA-Trans.

FATA-Trans: Field And Time-Aware Transformer for Sequential Tabular Data