Zhen Guo

Defeating the COVID-19 Infodemic on Twitter: A SIP Agent-Based Model of Rumor Propagation and Truth Bot Intervention

March 25, 2021   /  

Student Name: Zhen Guo
Majors: Computer Science, Sociology
Advisors: Thomas Tierney (Sociology), Heather Guanera (Computer Science)

Most Timely Award

As stated by a World Health Organization (WHO) director, “We’re not just fighting an epidemic [COVID-19]; we’re fighting an infodemic” (Zarocostas 2020). Researchers and policy makers have tried different methods to stop rumor propagation to defeat the infodemic. To help them, I built a virtual Twitter world, where computer-generated agents view hashtags and write tweets just like humans, to test effective policies. My virtual Twitter experiment gives a real-world implication that removal of the initial rumor spreaders on social media might not be necessary, since all the human agents end up believing in truth when rumor bots are still active. This study explores the intersection of fields including sociology, communication, network science, and computer science. Gabriel Tarde’s Law of Imitation and Bruno Latour’s Actor-Network Theory served as the theoretical frameworks of my project. I utilized agent-based model and SIR model to develop a new model called Spreader-Ignorant-Philosopher (SIP) model. It suggests a new way of understanding a complex social phenomenon using simplified computer agents. In the future, a more refined and realistic model might be used to test the effectiveness of real-world policies prior to its application in the real population.

Zhen will be online to field comments on April 16: 2-4 pm EDT (PST 11am-1pm, Africa/Europe: evening). Zhen will be available for a chat on Teams during Real-time chat session. Join a chat through this link.

55 thoughts on “Defeating the COVID-19 Infodemic on Twitter: A SIP Agent-Based Model of Rumor Propagation and Truth Bot Intervention”

  1. Zhen, I have missed seeing you and others this year. What a wonderful I.S. topic! I’m sure you did an amazing job. Congratulations!

    1. Thank you so much, Dr. Ramsay! I have missed seeing you as well. The experience I gained in AMRE definitely contributed to my IS. Thank you for your mentoring during my two years of AMRE project. I hope you have been enjoying your retirement.

  2. The spread of misinformation is such a fascinating concept. It’s exciting to see your research and conclusions on the use of truth bots to combat the spread. Congratulations on your IS!

  3. Modeling rumor propagation in social media is a timely and interesting research. Congratulations on this work and award, Zhen!

    1. Thank you Prof. Visa! And thank you for providing me the idea of building an agent-based model!

    1. Thank you Dr. Regan! I really appreciate your support and guidance in the sophomore research. My IS was built upon the ideas I learned from the Kilobots.

  4. Congratulations, Zhen! I wish you the very best in the future! You will do great things, I’m sure.

  5. I knew your general topic, but it’s very cool seeing the scope of the entirety. Congratulations on a job well done! With how interdisciplinary it was, I know how much work this must have involved. Your conclusions are really interesting, and I find particularly interesting your point that “rumors might be true” and that it could be dangerous to accidentally stifle them. Importantly, I think your implications here have great potential for being “accepted” by diverse sectors of society; broad acceptance is often such a hard thing to achieve, and yet the only way to make progress.

    “The light shines in the darkness, and the darkness can never extinguish it.” John 1:5

    1. Thank you, Carla! I really appreciate your support throughout my college year! And wow, I think you got
      the most important messages from my IS.

  6. Thank you for this presentation Zhen! The rapid rate of spread for misinformation is a very serious social and technical issue of our day. I like the approach that a “ban” or punishment would not be necessary for the majority of spreaders as it is most likely done due to a misunderstanding. “Curing” fiction via facts would most likely result in less resentment over being “silenced” on a social platform. How do you think that we could leverage this approach with the current methods that platforms like Twitter have taken to annotate tweets with contradictory or misleading information?

    Again, thank you for sharing this work, and great job!

    1. Hi Prof. Guarnera! I am glad your are here and I am happy to answer your question. I definitely agree that the truth bot method could be combined with the annotate tweets methods. Spreading true information could “immune” people from believing in rumors, but the range and scope of truth spreading might be limited due to the comprehensive nature of Twitter, like people might not view COVID-related hashtags and just read from their friends. In this case, labeling possible misinformation is a good idea.

  7. Congratulations! It is so awesome to see this all together after reading it in sections 🙂 Well done!!!

    1. Thank you Tessa! You are a great contributor to my success in IS. Without your support in writing, I can still finish my IS, but it would be very painful.

  8. Zhen, you have done fantastic work this year and I’m so proud of you. I have learned so much about sociology from you – between moral panic theory, actor network theory, and the juxtaposed ideas of Durkheim in contrast to Latour & Tarde. You have done an excellent job applying these theories to realistically simulate the spread of rumors through a social network, resulting in a topical and impressive IS. Congratulations!

    1. Thank you Prof. Guarnera! Sorry that your first year in Wooster had to be online… I am so proud to be one of your first IS students, and I hope you would enjoy the time in Wooster!

  9. Congratulations, Guo Zhen !!! Amazing interdisciplinary topic!!! I am so proud of you!

  10. You know I always say this to you but your ability to bring together sociology and computer science into an important and timely project is absolutely remarkable. I loved learning about this project and can’t wait to see where your career takes you in the future.

    1. Thank you Prof. Nurse! I applied a lot of statistical techniques learned from your class. I also really appreciate the way of thinking you have taught me.

  11. Congratulations, Zhen, for completing the most timely research! I appreciate the policy implications of your research, and look forward to hearing more great achievements from you in the future!

  12. Great slides, Zhen, although I think I should be Tweeting my congratulations, if I only knew how! I thoroughly enjoyed working with you and Prof. Guarnera on this project, and am so impressed with how willing you were to dive into unfamiliar theoretical material, work through the complicated modelling process, and then pull it all together into a compelling and important thesis. Most of all I appreciated your (and Prof. Guarnera’s) graciousness in handling my uninformed questions about modelling, and off-the-wall suggestions, like repeated invocation of the “rhizome” imagery. It was a genuine pleasure!

    1. Thank you, Prof. Tierney! I would not be able to complete this as complex as it is now without your guidance on the theoretical framework. It was a great pleasure to work with you, and I feel so honored to be one of your advisees.

  13. Congratulations my lovely (pre-) roomie!!! it is so cool to see how you integrated different diciplent of sociology and CS well and cooperate it with the most up-to-date issue! I am very proud of you! my questions are: are you planning on continueing you IS project in the future? What is the biggest take away message from doing IS for your life, besides what you have in your conclusion?

    1. Hi Coco, I would say making my model more realistic would be the next step. Currently, humans in my model can only transition among three statuses (I-S-P), they cannot be both at the same time. But in real life, a human might not spread a rumor even after being exposed to a lot of misinformation-this person might have a very critical mind. My biggest takeaway beyond my conclusions would be that rumor spreading and truth spreading are not theoretically different! There is a risk that a rumor becomes truth and truth becomes a rumor. Therefore, it is very important to be critical.

  14. I am proud of you, favorite roommate. It is interesting how you said as time goes on, the truth will reveal itself. I wonder how long this will take. I wonder rumor spreading could work for other areas in politics, as many people only seem to believe in the misinformation even the truths are revealed later.

    1. That’s right, one of the implications of my model is that policymakers can spread “truth” in advance to immune people from believing in rumors.

  15. Wow, what a creative way to think about an infodemic on Twittosphere! Thanks for sharing, Zhen. As you are aware, we use the metric R0 (R-naught) to study the transmissibility of an epidemic. Your thesis makes me wonder whether a similar metric can be developed for the rapidity of spread of misinformation in an extended Twitter network – perhaps the agent-based and SIR modeling framework you deploy already deploys such metrics. I would love to hear your POV.

    1. You are absolutely right! I am glad to see you pointing out the SIR disease spreading model can be applied to social phynomenons.

  16. Wow inspiring work Zhen! I am so impressed with how well you integrated both your majors in your thesis. I am glad to hear that the truth prevails in the end, at least in this simulation! Great job!

  17. Fantastic job, Zhen. It’s been a pleasure working with you, and I wish you all the best for your future! Blessings!

  18. This is terrific, Zhen! Congratulations on the award, and we wish you all the best in your pursuits after graduation!

    1. Thank you, Dr. Pasteur! I missed seeing you in person and really appreciate your guidance during AMRE and throughout my college years!

  19. So impressed with the usefulness and sophistication of yours creative research and model. Impresiive work. Fran Hall (Matt’s grandmother)

  20. This your work on this topic is so important given the rise of disinformation and its impact on public health. The impact of a pandemic is determined by more than just the physiological effects of the disease, but by people’s behavior in response. I hope you continue to investigate solutions to these issues!

Comments are closed.