News Recommender Experimental Research Infrastructure Survey
This is a quick survey to collect research questions from people who would use a proposed experimental news research infrastructure. The proposed infrastructure would include at least:
* a collection of users reading news through an instrumented app or web interface
* a mechanism for conducting experiments that vary the news recommender
* a mechanism for conducting experiments that vary the news interface
* support for both shorter- and longer-term field experiments
* support for pop-up experiments (including survey components)
* ways of monitoring longer-term usage of users in experimental groups
The intent is to support a wide range of research, including recommender systems, natural language processing, information retrieval, machine learning, human-computer interaction, and research at the intersection of computing with journalism, social science, and other fields.
Please fill in one research project per survey; we invite you to submit multiple surveys if that would make it easier to separate different ideas.
Research Specifics Intro
The following questions request detail on your research interests and plans
What field(s) of research does this question involve?
Natural Language Processing
High Level Study/Experiment Overview
Citations (paste a DOI, link, or citation) of 3-5 most relevant articles including: your published work in the area, prior research that leads to the research questions, and or other useful references if writing a short synopsis of this potential use.
Number of users desired?
How long would you intervene with these users?
How long would you want to follow their activity to track/measure impact?
What would be manipulated here (select all that apply?)
recommender / selection algorithm
auxiliary algorithms: e.g., summarization, NLP algorithms, explanation generation,
interface elements as it relates to browsing/searching/finding news content
overall interface (e.g., providing a voice interaction to replace visual)
interface for an individual article being read
content of the article itself
social features (network, sharing, liking etc.)
Are there specific features in such an infrastructure that would be critical to your research effort? If so, please specify:
Never submit passwords through Google Forms.
This form was created inside of University of Minnesota Twin Cities.