FormLM: Recommending Creation Ideas for Online Forms by Modelling Semantic and Structural Information

TL;DR

FormLM is a novel model that integrates semantic and structural information to assist in designing online forms, significantly improving question and block type recommendations based on a large new dataset.

Contribution

This work introduces FormLM, the first model to incorporate form structure into language modeling for form creation, and provides a new dataset of 62K online forms for training and evaluation.

Findings

FormLM outperforms general language models on form-related tasks.

Achieved 4.71 ROUGE-1 improvement in question recommendation.

Achieved 10.6 Macro-F1 improvement in block type suggestion.

Abstract

Online forms are widely used to collect data from human and have a multi-billion market. Many software products provide online services for creating semi-structured forms where questions and descriptions are organized by pre-defined structures. However, the design and creation process of forms is still tedious and requires expert knowledge. To assist form designers, in this work we present FormLM to model online forms (by enhancing pre-trained language model with form structural information) and recommend form creation ideas (including question / options recommendations and block type suggestion). For model training and evaluation, we collect the first public online form dataset with 62K online forms. Experiment results show that FormLM significantly outperforms general-purpose language models on all tasks, with an improvement by 4.71 on Question Recommendation and 10.6 on Block Type Suggestion in terms of ROUGE-1 and Macro-F1, respectively.