Naman Awasthi

One of the central difficulties of addressing the COVID-19 pandemic has been accurately measuring and predicting the spread of infections. In particular, official COVID-19 case counts in the United States are under counts of actual caseloads due to the absence of universal testing policies. Researchers have proposed a variety of methods for recovering true caseloads, often through the estimation of statistical models on more reliable measures, such as death and hospitalization counts, positivity rates, and demographics. However, given the disproportionate impact of COVID-19 on marginalized racial, ethnic, and socioeconomic groups, it is important to consider potential unintended effects of case correction methods on these groups. Thus, we investigate two of these correction methods for their impact on a downstream COVID-19 case prediction task. For that purpose, we tailor an auditing approach and evaluation protocol to analyze the fairness of the COVID-19 prediction task by measuring the difference in model performance between majority-White counties and majority-minority counties. We find that one of the correction methods improves fairness, decreasing differences in performance between majority-White and majority-minority counties, while the other method increases differences, introducing bias. While these results are mixed, it is evident that correction methods have the potential to exacerbate existing biases in COVID-19 case data and in downstream prediction tasks. Researchers planning to develop or use case correction methods must be careful to consider negative effects on marginalized groups.

We discuss MET, a learning-based algorithm proposed for perceiving a patient's level of engagement during telehealth sessions. We leverage latent vectors corresponding to Affective and Cognitive features frequently used in psychology literature to understand a person's level of engagement in a semi-supervised GAN-based framework. We showcase the efficacy of this method from the perspective of mental health and more specifically how this can be leveraged for a better understanding of patient engagement during telemental health sessions. To further the development of similar technologies that can be useful for telehealth, we also plan to release a dataset MEDICA containing 1299 video clips, each 3 seconds long and show experiments on the same. Our framework reports a 40% improvement in RMSE (Root Mean Squared Error) over state-of-the-art methods for engagement estimation. In our real-world tests, we also observed positive correlations between the working alliance inventory scores reported by psychotherapists. This indicates the potential of the proposed model to present patient engagement estimations that aligns well with the engagement measures used by psychotherapists.

Place-based short-term crime prediction models leverage the spatio-temporal patterns of historical crimes to predict aggregate volumes of crime incidents at specific locations over time. Under the umbrella of the crime opportunity theory, that suggests that human mobility can play a role in crime generation, increasing attention has been paid to the predictive power of human mobility in place-based short-term crime models. Researchers have used call detail records (CDR), data from location-based services such as Foursquare or from social media to characterize human mobility; and have shown that mobility metrics, together with historical crime data, can improve short-term crime prediction accuracy. In this paper, we propose to use a publicly available fine-grained human mobility dataset from a location intelligence company to explore the effects of human mobility features on short-term crime prediction. For that purpose, we conduct a comprehensive evaluation across multiple cities with diverse demographic characteristics, different types of crimes and various deep learning models; and we show that adding human mobility flow features to historical crimes can improve the F1 scores for a variety of neural crime prediction models across cities and types of crimes, with improvements ranging from 2% to 7%. Our analysis also shows that some neural architectures can slightly improve the crime prediction performance when compared to non-neural regression models by at most 2%.