Personalize My Story
Automatically Adapting News Article Text for Individual Users
Algorithmic news curation aggregators (e.g. Google News) are sometimes known to personalize the selection of stories shown to individuals. But far less is known about the potential for article-level personalization in which an article is automatically re-written to appeal to different types of users, perhaps even adapted to each individual. Could this be used to manipulate, persuade, inform, or engage people more effectively? The goal of this project is to prototype one or more templates for automated news articles that adapt to different types of people or individuals based on a given user model based on the types of information a news site might know (e.g. gender, age, race, location, interest-level, etc.). These templates will be used to produce personalizable news articles that are published to the web.
Faculty and Staff Leads
Project Details
2019 Winter
Important Questions
- What is an appropriate article topic with available data for which a story could be meaningfully personalized in a variety of ways?
- What dimensions of a user model are available and practical to use to adapt a news article?
- How can dynamic and data-driven templates be written using a tool like Arria Studio to produce different compelling versions of that story based on that user model?
- What are the ethical issues news organizations should be aware of in deploying this type of personalized article approach?
Sample Milestones
- Week 1-3 background research on user models, topic for story, data availability, ethical issues
- Week 4-5 write pilot story templates in Arria Studio that adapt to different facets of user model
- Week 6-7 refinement of story templates based on user feedback; data flow and system development to produce articles for publication
- Week 8-9 create project website where users can interactively provide a user model and see output of automatically generated article; refinement and publication of website
Outcome
The development of one or more data-driven templates that use a user model to automatically generate personalized news articles that are published direct to the web.
Links
- Yixue Wang and Nicholas Diakopoulos. Considerations for Article-Level Personalization of News Content. Algorithmic Personalization and News (APEN) Workshop. 2018.
- Adar, E. et al., 2017. PersaLog: Personalization of News Article Content. In Proc. CHI.
- Ten effective ways to personalize news platforms. June, 2018.
- Five risks of news personalization. June, 2018.
- Leppänen, L., Munezero, M. & Sirén-Heikel, S., 2017. Finding and Expressing News From Structured Data. In Academic Mindtrek.
Students
