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

Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 8133 - 8149 December 7-11, 2022 ©2022 Association for Computational Linguistics FormLM: Recommending Creation Ideas for Online Forms by Modelling Semantic and Structural Information Yijia Shao1∗ Mengyu Zhou2† Yifan Zhong3* Tao Wu4 Hongwei Han5* Shi Han2 Gideon Huang4 Dongmei Zhang2 1Peking University 2Microsoft Research 3Fudan University 4Microsoft 5Tsinghua University shaoyj@pku.edu.cn, {mezho, twu, shihan, gihuang, dongmeiz}@microsoft.com yfzhong20@fudan.edu.cn, hhw20@mails.tsinghua.edu.cn Abstract Online forms are widely used to collect data from human and have a multi-billion market. Many software products provide online ser- vices 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 de- signers, in this work we present FormLM to model online forms (by enhancing pre-trained language model with form structural informa- tion) and recommend form creation ideas (in- cluding question / options recommendations and block type suggestion). For model training and evaluation, we collect the first public online form dataset with 62K online forms. Experi- ment 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. 1 Introduction Online forms are widely used to collect data in ev- eryday scenarios such as feedback gathering (Ilieva et al., 2002), application system (Sylva and Mol, 2009), research surveys (Yarmak, 2017), etc. With a multi-billion market (Research and Markets, 2021), many software products – such as Survey Monkey (Abd Halim et al., 2018), Google (Mondal et al., 2018) and Microsoft Forms (Rhodes, 2019) – provide services to help users create online forms which consist of multiple blocks (e.g., Figure 1). However, there are obstacles preventing the cre- ation of well-designed online forms, which could hurt response rate and quality (Krosnick, 2018). For each form question, form designers need to ∗ The contributions by Yijia Shao, Yifan Zhong and Hongwei Han have been conducted and completed during their internships at Microsoft Research Asia, Beijing, China. † Corresponding author. Option 1 Choice Add option Team Building Questionnaire Please fill out this form to allow us to better understand our employees‘ interests, strengths, and real feelings. Suggested: Employee ID Your answer Text Field Full Name Your answer Question Recommendation Options Recommendation How happy are you with your current job? Block Type Suggestion Suggested options: Add all | Yes No Do you feel you and your manager get along? Add rating score Rating TODO Figure 1: An Example Online Form with the Three Tasks of Intelligent Form Creation Ideas. write an informative title, specify its type, and pro- vide other required components. This process is tedious and time-consuming even for experienced users. Also, non-experts may be unsure about what question to add or which question type to choose. To improve the experience and efficiency of form composing, it is desirable that online form services could recommend creation ideas to …

Content truncated to 500 words