From Knoesis wiki
Jump to: navigation, search

COVID-19 Research Services at AIISC

COVID-19: Public Health Study: Semantic Analysis of Social Media and New Big Data to understanding COVID-19's impact on mental health, addiction and gender-based violence


Experts have warned about the potential rapid growth in several social and health consequences of COVID-19 on individuals and society, specifically Mental Health (Depression, Anxiety), Addiction (Substance-use), and Gender-based (or Domestic) Violence (GBV). We have been successfully utilizing social media measures for epidemiology and public health research, such as Drug abuse (leading to FDA warning), Mental health, harassment, and GBV. For COVID-19 we are exploring the following questions:

  1. Q1: How can we use social media to measure psychological and social impact in (near) real-time?
  2. Q2: Specifically, how does intervention in the form of state-level policy choices and implementations relate to mental health and addiction-related behaviors across different states? What evidence is there for adaptive/coping behavior?
  3. Q3: How do GenZ and Millennials express themselves in the outbreak, particularly in the context of Mental health?

Our approach for processing big social media data involves a series of state-of-the-art AI techniques utilizing human-curated knowledge bases, data mining, and semantic filtering procedures (see Technical Approach for details). We have collected >800 Million tweets from March 14 to April 10, 2020, and subsequently selecting 45 M Tweets with location (explicit location), 27 M Tweets of those with matches to entities in News, finally obtaining 15 M Tweets with exact matches to concepts in our human-curated Mental Health and Drug Abuse computationally accessible Knowledge Base (MHDA-Kb). Figure 1 illustrates the links between the words in tweets and their mental health interpretation in this knowledge base. The interpretation of social media data is assisted through ~700K COVID-related news articles (January 01 to March 29, 2020).

Examples tweets include that motivated our questions:

  1. "You believe I have any pleasure in this chaos? Jeez. I’ve been despairing for 2 months."
  2. "A feeling of hopelessness. Seems I am in a dark age. #coronavirus #COVID19".
  3. “self-isolated for two weeks and depression becoming unbearable. This coronavirus is worsening my anxiety a lot and I am terrified.”
  4. “side effects of hydroxychloroquine: "Mental/mood changes (such as confusion, personality changes, unusual thoughts/behavior, depression, feeling being watched, hallucinating"
  5. “The reason I’m on #Hxychloroquine sedative is that I actually have anxiety in my sleep. I don’t know why I’m having nocturnal anxiety attacks. It’s fucking awful. I’ll wake up and not even know where I am because I’m so scared or shook”

Social Quality Index (SQI): A Social Quality Index (SQI) is calculated from the aggregation of mental health and addiction components. Raw SQI takes into account tweet concepts abstracted through three different mental health lenses in the MHDA-Kb: Depression, Anxiety, and Drug Abuse Disorders. Raw SQI simply aggregates the relevant features with respect to each of these lenses in each message, and does not take into account preceding state conditions. Change in SQI is also potentially informative, particularly for comparisons between states. We transformed raw state SQI into a relative state ranking, to capture drifts between worsening and improving psychological conditions in social quality. SQI ranking is also used to examine the effect of external factors, such as school closure, business closure, unemployment, and lockdown (including the extension of lockdown).

News Coverage

  • [1] We’re measuring online conversation to track the social and mental health issues surfacing during the coronavirus pandemic, The Conversation, 20 April 2020
  • [2] What Your Social Media Posts Say About Your Stress Level Right Now, Healthline, 30 April 2020

Other Relevant Research:

  • [3]Depressive, Drug Abusive, or Informative: Knowledge-aware Study of News Exposure during COVID-19 Outbreak


Dr. Amit P. Sheth

Founding Director, Artificial Intelligence Institute

The University of South Carolina

Dr. Valerie L. Shalin

Profession and Human Factors Area Leader

Department of Psychology

Wright State University


Principal Investigators: Prof. Amit P. Sheth

Co-Investigators: Prof. Valerie L. Shalin

Postdoctoral Researchers: Dr. Ugur Kursuncu

Graduate Students: Manas Gaur, Vedant Khandelwal, Vishal Pallagani, Usha Lokala

This research is funded in part by NSF Award "Spokes: MEDIUM: MIDWEST: Collaborative: Community-Driven Data Engineering for Substance Abuse Prevention in the Rural Midwest" (Award#: 1956009)

COVID19: Epidemiology Study with Exogenous Factors

[4]Study accepted at ACM KDD'20 AI for COVID Track

Epidemiological models are the mathematical models that capture the dynamics of epidemics. The spread of the virus has two routes - exogenous and endogenous. The exogenous spread is from outside the population under study, and endogenous spread is within the population under study. Although some of the models consider the exogenous source of infection, they have not studied the interplay between exogenous and endogenous spreads. In this paper, we introduce a novel model - the Exo-SIR model that captures both the exogenous and endogenous spread of the virus. We analyze to find out the relationship between endogenous and exogenous infections during the Covid19 pandemic. First, we simulate the Exo-SIR model without assuming any contact network for the population. Second, simulate it by assuming that the contact network is a scale-free network. Third, we implemented the Exo-SIR model on a real dataset regarding Covid19. We found that endogenous infection is influenced by even a minimal rate of exogenous infection. Also, we found that in the presence of exogenous infection, the endogenous infection peak becomes higher, and the peak occurs earlier. This means that if we consider our response to a pandemic like Covid19, we should be prepared for an earlier and higher number of cases than the SIR model suggests if there are the exogenous source(s) of infection.


Dr. Amit P. Sheth

Founding Director, Artificial Intelligence Institute

The University of South Carolina

Manas Gaur

Ph.D. Candidate, Artificial Intelligence Institute

The University of South Carolina

Collaborators: [5] Nirmal Sivaraman, [6]Dr. Sakthi Balan

Health-e Gamecock COVID-19 Daily Symptom Monitoring mHealth App/Chatbot

The ongoing rise of new infections with the novel coronavirus, COVID-19, presents a special set of challenges for colleges and universities as students and employees return to campus. Campuses represent a unique population and setting that require an innovative and novel solution to keeping students and employees safe and preventing the spread of the virus. The Health-e Gamecock COVID-19 Daily Symptom Monitoring mobile health application (mHealth app) with an integrated chatbot (i.e., a conversational agent that mimics human conversation) and the dashboard was built by researchers and clinicians in the College of Nursing (CON) and the Artificial Intelligence Institute (AIISC) at the University of South Carolina. The Health-e Gamecock mHealth app was developed using AIISC developed Health-e Gamecock platform incorporating prior mApp/Chatbot development efforts, research evidence, and healthcare expert opinions. It uses a comprehensive approach to understand the incidence and prevalence of certain physical (e.g., cough, difficulty breathing, loss of smell) and emotional/mental health (e.g., stress, anxiety, depression) symptoms that may be associated with COVID-19 infection. Additionally, it was built on the premise that monitoring symptoms throughout this time are critical to success for mitigating the spread of COVID-19 in addition to testing, wearing a face covering, social distancing, and handwashing The application is available for IoS or Android mobile platforms, and on the Web. It supports the collection of daily symptoms relevant to COVID-19 from any participant campus community member and provides access to relevant news, education, and training material to the participant. Aggregate and anonymized data thus collected give real-time monitoring of the health of the community selected- a college or the entire campus. This application only collects anonymized data—no personally identifiable information is collected. Industry-standard security measures are used encompassing mobile and cloud components.

IRB approved: research study Changes in COVID-19-Related Symptoms Across a College Campus Using a mHealth Application ( Pro00102203,06 Aug 2020)


Dr. Amit P. Sheth

Founding Director, Artificial Intelligence Institute

The University of South Carolina

Dr. Ronda G. Hughes

Director, Center for Nursing Leadership and Associate Professor

College of Nursing

The University of South Carolina

Covid19 Mask Analysis Program (CMAP)

COVID19 is a global pandemic whose impact in regions around the world has varied widely, as measured by the number of cases and deaths, depending on the local demographics as well as public health policies implemented in response, e.g., mask coverings. In this work, we present a tool called COVID Mask Analyzer Program (CMAP) that can be used to understand the impact of mask policies at local and national scale. Internally, the tool uses the well established techniques of robust synthetic control and New York Times' data about mask adherence and cases to answer counter-factual questions.

CMAP was developed in partnership between Tantiv4 and the PI at the AI Institute. It uses advanced data cleaning and normalization methods, and covers counties around the United States. This work opens up new avenues of research in human-machine collaboration to foster data-driven public health policies.

See the video to see the tool in action and read the demonstration paper for details. As an example of output, CMAP shows that for Richland county, SC, intervention by June 1, 2020 would have been most consequential in saving lives compared to July 1 or Aug 1.

News Coverage:

  • [7] USC researchers build model showing how many coronavirus infections masks prevent, Post & Courier, 10 Oct 2020
  • [8] A new data-driven model shows that wearing masks saves lives – and the earlier you start, the better, The Conversation, 13 Nov 2020


  1. [9] Sparsh Johri, Kartikaya Srivastava, Chinmayi Appajigowda, Lokesh Johri and Biplav Srivastava. "A Nation-Wide Tool To Understand Impact of COVID19 Related Mask Policies Using Robust Synthetic Control."

PI: Dr. Biplav Srivastava, AI Institute, University of South Carolina

Collaborators: Sparsh Johri, Kartikaya Srivastava, Chinmayi Appajigowda, Lokesh Johri

Relevant Articles/Publication on Research Used in the above work:

  1. Parth Asawa, Manas Gaur, Kaushik Roy, and Amit Sheth. "COVID-19 in Spain and India: Comparing Policy Implications by Analyzing Epidemiological and Social Media Data." Proceedings of AAAI 2020 Fall Symposium on AI for Social Good.
  2. Ugur Kursuncu, Manas Gaur, and Amit Sheth. "Knowledge infused learning (K-IL): Towards deep incorporation of knowledge in deep learning." Proceedings of the AAAI 2020 Spring Symposium on Combining Machine Learning and Knowledge Engineering in Practice (AAAI-MAKE 2020). Stanford University, Palo Alto, California, USA, March 23-25, 2020.
  3. Amit Sheth, Manas Gaur, Ugur Kursuncu, and Ruwan Wickramarachchi. "Shades of knowledge-infused learning for enhancing deep learning." IEEE Internet Computing 23, no. 6 (2019): 54-63.
  4. Manas Gaur, Amanuel Alambo, Joy Prakash Sain, Ugur Kursuncu, Krishnaprasad Thirunarayan, Ramakanth Kavuluru, Amit Sheth, Randy Welton, and Jyotishman Pathak. "Knowledge-aware assessment of severity of suicide risk for early intervention." In The World Wide Web Conference, pp. 514-525. 2019.
  5. Ugur Kursuncu, Manas Gaur, Carlos Castillo, Amanuel Alambo, Krishnaprasad Thirunarayan, Valerie Shalin, Dilshod Achilov, I. Budak Arpinar, and Amit Sheth. "Modeling islamist extremist communications on social media using contextual dimensions: Religion, ideology, and hate." Proceedings of the ACM on Human-Computer Interaction 3, no. CSCW (2019): 1-22.
  6. Ugur Kursuncu, Manas Gaur, Usha Lokala, Krishnaprasad Thirunarayan, Amit Sheth, and I. Budak Arpinar. "Predictive analysis on Twitter: Techniques and applications." In Emerging research challenges and opportunities in computational social network analysis and mining, pp. 67-104. Springer Nature, 2019.
  7. Manas Gaur, Ugur Kursuncu, Amanuel Alambo, Amit Sheth, Raminta Daniulaityte, Krishnaprasad Thirunarayan, and Jyotishman Pathak. "" Let Me Tell You About Your Mental Health!" Contextualized Classification of Reddit Posts to DSM-5 for Web-based Intervention." In Proceedings of the 27th ACM International Conference on Information and Knowledge Management, pp. 753-762. 2018.
  8. Ugur Kursuncu, Manas Gaur, Usha Lokala, Anurag Illendula, Krishnaprasad Thirunarayan, Raminta Daniulaityte, Amit Sheth, and I. Budak Arpinar. "What's ur Type? Contextualized Classification of User Types in Marijuana-Related Communications Using Compositional Multiview Embedding." In 2018 IEEE/WIC/ACM International Conference on Web Intelligence (WI), pp. 474-479. IEEE, 2018.
  9. Andrew J. Hampton, and Valerie L. Shalin. "Sentinels of breach: Lexical choice as a measure of urgency in social media." Human factors 59, no. 4 (2017): 505-519.
  10. Raminta Daniulaityte, Lu Chen, Francois R. Lamy, Robert G. Carlson, Krishnaprasad Thirunarayan, and Amit Sheth. "“When ‘bad’is ‘good’”: identifying personal communication and sentiment in drug-related tweets." JMIR public health and surveillance 2, no. 2 (2016): e162.


Related Projects