Table to Text Generation with Subjectivity and Objectivity

Anonymous

Abstract: Table-to-text generation, a long-standing challenge in natural language generation, has remained unexplored through the lens of subjectivity. Subjectivity here encompasses the comprehension of information derived from the table that cannot be described solely by objective data. To ascertain the relevance and social applications of this work, we conduct a public survey involving relevant people. The survey results unequivocally conclude the significance of this research. Given the absence of pre-existing datasets, we introduce the $\textbf{TaTS dataset}$. A new linearization technique is implemented to flatten the tables. We perform the task using various sequence-to-sequence models and analyze the results from a qualitative perspective to ensure the capture of subjectivity. To the best of our knowledge, this is the first kind of dataset on tables with subjectivity included.

Paper Type: short

Research Area: Generation

Contribution Types: Model analysis & interpretability, NLP engineering experiment, Data resources, Data analysis

Languages Studied: English

Consent To Share Submission Details: On behalf of all authors, we agree to the terms above to share our submission details.