Interested in being published? Check out these research opportunities!

Professor Jennifer Golbeck has some research opportunities that iSchool students can get involved with. Check the project descriptions below out and reach out to Professor Golbeck at golbeck@cs.umd.edu if anything piques your interest!

Behavioral biometrics

(Requires good programming skills)

We can often detect who an anonymous person is by looking at the traces of data they leave behind in their online activities and comparing it to data from known people. This project is looking at how to develop “behavioral biometrics “– measures of behavior that can help us identify a person – for deanonymizarion. I’m particularly interested in looking at traces of social interactions, but things like timing, language patterns, and other interaction data can be useful.

I have data sets from two different social networks that include time stamped interaction data between people and objects. If you are interested in working on this project, send me mail that describes your programming background (particularly working with formatted text data), and tell me one measurement that you would investigate in these data sets to see if it can possibly link an unknown person to a known profile. Your answer doesn’t have to be “right” (this is research so I don’t even know what “right” would be), but I want to see how you think about this problem. You have to be self-motivated and directed on this, since we’re mostly exploring the space.

Trollbusting

(No technical background required)

This is a continuation of a project we started over the summer where we are analyzing a large volume of comments and labeling them as trolling or not. We didn’t finish the work over the summer so I’m recruiting a new batch of students to help us build this data set, which I believe will be a very important contribution to research on online harassment. The first set of work is simply for you to read tweets and label them according to our scheme. We will train you on how to apply the labels and you simply go to a website, read the comments, and select the appropriate labels. This can be done anytime and all I ask is about one hour week as a minimum commitment from each person on this project. Anyone who puts in that amount of time will be included as a co-author on the paper we publish (hopefully soon).

Once we get the labeling complete, there will be work for people with computational linguistics experience to work on automatically detecting  the trolling comments.

Bonus: you get an awesome free t-shirt as part of this team.

Cancer patient timelines

(No Technical background required)

I’m in the process of developing a project that hopefully will receive funding in the future, but for now we’re in the data collection and initial analysis stage. I’m interested in examining the public social media timelines of cancer patients from diagnosis until recovery or death. We are looking for critical moments and that timeline when people are more or less active, when they could use more social support, and how that manifests in the social media timelines.

In this early stage, it means searching for these people online and building a collection of usernames and comments for patients that recover and those who do not. People working on this project would help develop the methodology for finding these patients across different social media platforms and work on collecting the data. Once we have a reasonably sized set to start with, we will read those posts and develop the first set of important markers in people’s progression through the disease. Eventually, I hope to grow this into a large project that incorporates elements of computational linguistics, social network analysis, and other technologies to build supportive interfaces and insights that can help patients, physicians, and their support networks.

Leave a Reply

Your email address will not be published. Required fields are marked *

*