Eliciting New Wikipedia Users' Interests via Automatically Mined Questionnaires: For a Warm Welcome, Not a Cold Start

Every day, thousands of users sign up as new Wikipedia contributors. Once\njoined, these users have to decide which articles to contribute to, which users\nto seek out and learn from or collaborate with, etc. Any such task is a hard\nand potentially frustrating one given the sheer size of Wikipedia. Supporting\nnewcomers in their first steps by recommending articles they would enjoy\nediting or editors they would enjoy collaborating with is thus a promising\nroute toward converting them into long-term contributors. Standard recommender\nsystems, however, rely on users' histories of previous interactions with the\nplatform. As such, these systems cannot make high-quality recommendations to\nnewcomers without any previous interactions -- the so-called cold-start\nproblem. The present paper addresses the cold-start problem on Wikipedia by\ndeveloping a method for automatically building short questionnaires that, when\ncompleted by a newly registered Wikipedia user, can be used for a variety of\npurposes, including article recommendations that can help new editors get\nstarted. Our questionnaires are constructed based on the text of Wikipedia\narticles as well as the history of contributions by the already onboarded\nWikipedia editors. We assess the quality of our questionnaire-based\nrecommendations in an offline evaluation using historical data, as well as an\nonline evaluation with hundreds of real Wikipedia newcomers, concluding that\nour method provides cohesive, human-readable questions that perform well\nagainst several baselines. By addressing the cold-start problem, this work can\nhelp with the sustainable growth and maintenance of Wikipedia's diverse editor\ncommunity.\n