Abstract
Social media sites are challenged by both the scale and variety of deviant behavior online. While algorithms can detect spam and obscenity, behaviors that break community guidelines on some sites are difficult because they have multimodal subtleties (images and/or text). Identifying these posts is often regulated to a few moderators. In this paper, we develop a deep learning classifier that jointly models textual and visual characteristics of pro-eating disorder content that violates community guidelines. Using a million Tumblr photo posts, our classifier discovers deviant content efficiently while also maintaining high recall (85%). Our approach uses human sensitivity throughout to guide the creation, curation, and understanding of this approach to challenging, deviant content. We discuss how automation might impact community moderation, and the ethical and social obligations of this area.
Original language | English (US) |
---|---|
Title of host publication | CHI 2017 - Proceedings of the 2017 ACM SIGCHI Conference on Human Factors in Computing Systems |
Subtitle of host publication | Explore, Innovate, Inspire |
Publisher | Association for Computing Machinery |
Pages | 3213-3226 |
Number of pages | 14 |
ISBN (Electronic) | 9781450346559 |
DOIs | |
State | Published - May 2 2017 |
Externally published | Yes |
Event | 2017 ACM SIGCHI Conference on Human Factors in Computing Systems, CHI 2017 - Denver, United States Duration: May 6 2017 → May 11 2017 |
Publication series
Name | Conference on Human Factors in Computing Systems - Proceedings |
---|---|
Volume | 2017-May |
Other
Other | 2017 ACM SIGCHI Conference on Human Factors in Computing Systems, CHI 2017 |
---|---|
Country/Territory | United States |
City | Denver |
Period | 5/6/17 → 5/11/17 |
Bibliographical note
Publisher Copyright:© 2017 ACM.
Keywords
- Computer vision
- Content moderation
- Deep learning
- Deviant behavior
- Pro-eating disorder
- Social media
- Tumblr