2022-07-15: Disinformation Detection and Analytics REU Program - Mid Summer Presentations


The Research Experiences for Undergraduates (REU) program provides hands-on learning and research experiences to undergraduates focused on disinformation detection and analytics. This 10-week summer program started on June 6th and it is facilitated at Old Dominion University (ODU) with the Web Science and Digital Libraries (WSDL) Research Group within the Department of Computer Science in partnership with the Virginia Modeling, Analysis, and Simulation Center (VMASC) and the Department of Communication Disorders and Special Education. The REU program is funded by the 2022 National Science Foundation (NSF) Research Experiences for Undergraduate Site. There are eight undergraduate students in this program doing projects on disinformation detection and analytics.


The REU project mid-summer presentations were held on Friday, July 8th at the Department of Computer Science. The undergraduates of the ODU CS REU program presented their projects, progress, and plans for the next five weeks. 

Mentors, Mentees, and Mentor Assistants of ODU CS REU Program

1. Caleb Bradford (@calebkbradford) advised by Dr. Michael Nelson (@phonedude_mln)

The REU mid-summer presentation session started with Caleb’s presentation on "Did they really tweet that?". In his project, Caleb tries to establish the probability of what public figures said, did not say, or said and then deleted. He worked on automating the process of querying the Politwoops fact-checking website for evidence and he continues to work on refining Politwoops search queries, querying with other fact-checking sites, researching about searching archives for evidence, and the verification.

2. Haley Bragg (@haleybragg17) advised by Dr. Michele C. Weigle (@weiglemc)

Haley presented her work on "Discovering the Traces of Disinformation on Instagram" which is an extension of Himarsha Jayanetti’s work on how well Instagram pages are covered in public web archives. Haley worked on developing methods for gathering metadata from CrowdTangle API and Memento collection and is currently working on expanding this research further.

3. Ash Dobrenen (@Kat9951) advised by Dr. Vikas Ashok (@vikas_daveb)

Ash presented "Protecting Blind Screen-Reader Users From Deceptive Content", in which they study on detecting and alerting blind users about deceptive content on web pages. After identifying the features for the machine learning model for detecting deceptive contents, they are now working on building a machine learning algorithm and evaluating the model’s performance. 

4. Michael Husk (@mhusk_3) advised by Dr. Faryaneh Poursardar (@Faryane)

In the mid-summer presentations, Michael presented "Fake Review Detection". In this work, he is trying to identify fake reviews posted on review websites such as Amazon. After achieving promising results with an accuracy of ~ 80% with the Multinominal Naive Bayes model on the Amazon review dataset, Michael continues to work with other ML models such as Bernoulli Naive Bayes, Logistic Regression, Random Forests, and BERT.

5. Dani Graber (@compsci_dani) advised by Dr. Anne Perrotti (@slpmichalek)

Dani presented “Disinformation About Mental Health on Tiktok”, in which she discussed how disinformation could lead to improper self-diagnosis among users. For this, she presented symptoms of mental health issues posted on social media with user comments highlighting possible self-misdiagnosis followed by the proper symptoms and the diagnosis process. She also discussed “why people want” mental health issues on Tiktok, highlighting potential fame, fans, and monetary benefits. 

6. Ethan Landers (@ethanlanders_) advised by Dr. Jian Wu (@fanchyna)

Ethan’s presentation was titled “An Assessment of Scientific Claim Verification Frameworks,” which attempts to detect misinformation in the scientific community, which can often be hard to distinguish. To achieve this, he is working on creating a generalized dataset containing scientific articles with some containing disinformation. Then he plans on using the dataset to analyze the results of scientific claim verification models. 

7. Isabelle Valdes (@isabellefv_) advised by Dr. Erika Frydenlund (@ErikaFrydenlund)

In the following presentation, Isabelle presented “Networks of Disinformation: Colombian Twitter and the Venezuelan Migration Crisis.” In the study, she analyzes the spread of disinformation affecting the migrants focusing on Venezuelan refugees, such as them being responsible for crimes. She attempts to identify networks that spread disinformation using graph cluster algorithms and interactions among social media users. 

8. Autumn Woodson (@AWoods_n) advised by Dr. Sampath Jayarathna (@OpenMaze)

As the final REU mid-summer presentation, Autumn presented her study on “Human Interaction With Fake News.” In this work, she analyzes an eye movement dataset to characterize reading patterns using eye-tracking while interacting with fake news articles. This study extends an eye-tracking study published at ETRA 2022 by combining pupillometric information.

The presentations offered a wide range of perspectives on different forms of disinformation in social media and their potential to confuse and harm users. Despite disinformation detection and analytics being a rapidly growing research area, the presentations highlighted many unexplored research avenues. 

The undergraduates will continue their projects until the end of the REU program on August 5th. We invite you to watch the WS-DL blog and follow @WebSciDL and @oducsreu on Twitter for more information.

-- Yasasi Abeysinghe (@Yasasi_Abey) & Bhanuka Mahanama (@mahanama94)

Comments