SoMeR: Multi-View User Representation Learning for Social Media
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909546345136128 |
|---|---|
| author | Guo, Siyi Burghardt, Keith Pantè, Valeria Lerman, Kristina |
| author_facet | Guo, Siyi Burghardt, Keith Pantè, Valeria Lerman, Kristina |
| contents | Social media user representation learning aims to capture user preferences, interests, and behaviors in low-dimensional vector representations. These representations are critical to a range of social problems, including predicting user behaviors and detecting inauthentic accounts. However, existing methods are either designed for commercial applications, or rely on specific features like text contents, activity patterns, or platform metadata, failing to holistically model user behavior across different modalities. To address these limitations, we propose SoMeR, a Social Media user Representation learning framework that incorporates temporal activities, text contents, profile information, and network interactions to learn comprehensive user portraits. SoMeR encodes user post streams as sequences of time-stamped textual features, uses transformers to embed this along with profile data, and jointly trains with link prediction and contrastive learning objectives to capture user similarity. We demonstrate SoMeR's versatility through three applications: 1) Identifying information operation driver accounts, 2) Measuring online polarization after major events, and 3) Predicting future user participation in Reddit hate communities. SoMeR provides new solutions to better understand user behavior in the socio-political domains, enabling more informed decisions and interventions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_05275 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SoMeR: Multi-View User Representation Learning for Social Media Guo, Siyi Burghardt, Keith Pantè, Valeria Lerman, Kristina Social and Information Networks Artificial Intelligence Information Retrieval Social media user representation learning aims to capture user preferences, interests, and behaviors in low-dimensional vector representations. These representations are critical to a range of social problems, including predicting user behaviors and detecting inauthentic accounts. However, existing methods are either designed for commercial applications, or rely on specific features like text contents, activity patterns, or platform metadata, failing to holistically model user behavior across different modalities. To address these limitations, we propose SoMeR, a Social Media user Representation learning framework that incorporates temporal activities, text contents, profile information, and network interactions to learn comprehensive user portraits. SoMeR encodes user post streams as sequences of time-stamped textual features, uses transformers to embed this along with profile data, and jointly trains with link prediction and contrastive learning objectives to capture user similarity. We demonstrate SoMeR's versatility through three applications: 1) Identifying information operation driver accounts, 2) Measuring online polarization after major events, and 3) Predicting future user participation in Reddit hate communities. SoMeR provides new solutions to better understand user behavior in the socio-political domains, enabling more informed decisions and interventions. |
| title | SoMeR: Multi-View User Representation Learning for Social Media |
| topic | Social and Information Networks Artificial Intelligence Information Retrieval |
| url | https://arxiv.org/abs/2405.05275 |