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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Hybrid Question-Answering (HQA), which targets reasoning over tables andpassages linked from table cells, has witnessed significant research in recentyears. A common challenge in HQA and other passage-table QA datasets is that itis generally unrealistic to iterate over all table rows, columns, and linkedpassages to retrieve evidence. Such a challenge made it difficult for previousstudies to show their reasoning ability in retrieving answers. To bridge thisgap, we propose a novel Table-alignment-based Cell-selection and Reasoningmodel (TACR) for hybrid text and table QA, evaluated on the HybridQA andWikiTableQuestions datasets. In evidence retrieval, we design atable-question-alignment enhanced cell-selection method to retrievefine-grained evidence. In answer reasoning, we incorporate a QA module thattreats the row containing selected cells as context. Experimental results overthe HybridQA and WikiTableQuestions (WTQ) datasets show that TACR achievesstate-of-the-art results on cell selection and outperforms fine-grainedevidence retrieval baselines on HybridQA, while achieving competitiveperformance on WTQ. We also conducted a detailed analysis to demonstrate thatbeing able to align questions to tables in the cell-selection stage can resultin important gains from experiments of over 90\% table row and column selectionaccuracy, meanwhile also improving output explainability.

TACR: A Table-alignment-based Cell-selection and Reasoning Model for Hybrid Question-Answering