Data Valuation with Gradient Similarity.

High-quality data is crucial for accurate machine learning and actionable analytics, however, mislabeled or noisy data is a common problem in many domains. Distinguishing low- from high-quality data can be challenging, often requiring expert knowledge and considerable manual intervention. Data Valuation algorithms are a class of methods that seek to quantify the value of each sample in a dataset based on its contribution or importance to a given predictive task. These data values have shown an impressive ability to identify mislabeled observations, and filtering low-value data can boost machine learning performance. In this work, we present a simple alternative to existing methods, termed Data Valuation with Gradient Similarity (DVGS). This approach can be easily applied to any gradient descent learning algorithm, scales well to large datasets, and performs comparably or better than baseline valuation methods for tasks such as corrupted label discovery and noise quantification. We evaluate the DVGS method on tabular, image and RNA expression datasets to show the effectiveness of the method across domains. Our approach has the ability to rapidly and accurately identify low-quality data, which can reduce the need for expert knowledge and manual intervention in data cleaning tasks.

Graph Structured Neural Networks for Perturbation Biology.

Computational modeling of perturbation biology identifies relationships between molecular elements and cellular response, and an accurate understanding of these systems will support the full realization of precision medicine. Traditional deep learning, while often accurate in predicting response, is unlikely to capture the true sequence of involved molecular interactions. Our work is motivated by two assumptions: 1) Methods that encourage mechanistic prediction logic are likely to be more trustworthy, and 2) problem-specific algorithms are likely to outperform generic algorithms. We present an alternative to Graph Neural Networks (GNNs) termed Graph Structured Neural Networks (GSNN), which uses cell signaling knowledge, encoded as a graph data structure, to add inductive biases to deep learning. We apply our method to perturbation biology using the LINCS L1000 dataset and literature-curated molecular interactions. We demonstrate that GSNNs outperform baseline algorithms in several prediction tasks, including 1) perturbed expression, 2) cell viability of drug combinations, and 3) disease-specific drug prioritization. We also present a method called GSNNExplainer to explain GSNN predictions in a biologically interpretable form. This work has broad application in basic biological research and pre-clincal drug repurposing. Further refinement of these methods may produce trustworthy models of drug response suitable for use as clinical decision aids. Availability and implementation: Our implementation of the GSNN method is available at https://github.com/nathanieljevans/GSNN. All data used in this work is publicly available.

Leveraging influencers to increase HPV vaccination intention: The impact of message framing and health regulatory fit using repeated measures.

Through a series of two online experiments that incorporate a repeated measures design, we examine how message framing (loss versus gain) within Instagram influencers' posts interact with consumers' health regulatory orientation (promotion versus prevention) to impact HPV vaccination intention between two data collection points. Findings indicate that among those who are more prevention oriented, exposure to a loss-framed influencer advertisement was effective at increasing intention to receive an HPV vaccine relative to those that were exposed to gain-framed influencer advertisement. Based on these findings we offer theoretical and managerial implications.

Using immersive virtual reality to improve the beliefs and intentions of influenza vaccine avoidant 18-to-49-year-olds: Considerations, effects, and lessons learned.

OBJECTIVE: Only one-third of adults 18-49 years old in the United States receive a recommended annual influenza vaccination. This study examined whether supplementing vaccine information statements (VIS) with an immersive virtual reality (VR), short video or electronic pamphlet story designed to convey the community immunity benefits of influenza vaccination would improve influenza vaccine avoidant participants' influenza-related perceptions as well as their influenza vaccination-related beliefs, confidence and intentions. METHOD: A one-way between-subjects experimental design compared the effects of adding a supplemental education experience prior to VIS exposure with flu vaccine avoidant 18-to-49-year-olds. The 171 participants recruited from the community were randomly assigned to one of three modality treatment conditions [VR, video, or e-pamphlet (i.e., story board presented via electronic tablet)] or a VIS-only control condition. RESULTS: Compared to the modalities, the VR intervention created a stronger perception of presence (i.e., feeling of "being there" in the story), which, in turn, increased participants' concern about transmitting influenza to others and raised vaccination intention. Increased concern about transmitting influenza to others was associated with positive effects on influenza vaccination-related beliefs, including confidence that one's flu vaccination would protect others. Neither the e-pamphlet nor the video intervention were able to elicit a sense of presence nor were they able to improve the impact of the VIS on the outcome measures. CONCLUSIONS: Immersive VR has much potential to increase understanding of key immunization concepts, such as community immunity, through creative executions that increase a sense of presence. Given the need to increase influenza vaccination uptake among 18-to-49-year-olds, and the projected growth in VR accessibility and use, additional applications and assessments related to vaccination communication and education are needed and warranted. By increasing the ability to convey key vaccine and immunization concepts, immersive VR could help address vaccination hesitancy and acceptance challenges.

It's Complicated: The 2014-2015 U.S. Measles Outbreak and Parents' Vaccination Beliefs, Confidence, and Intentions.

While it seems intuitive that highly visible vaccine-preventable disease outbreaks should impact perceptions of disease risk and facilitate vaccination, few empirical studies exist to confirm or dispel these beliefs. This study investigates the impact of the 2014-2015 Disneyland measles outbreak on parents' vaccination attitudes and future vaccination intentions. The analysis relies on a pair of public opinion surveys of American parents with at least one child under the age of six (N = 1,000 across each survey). Controlling for basic demographics, we found higher levels of reported confidence in the safety and efficacy of childhood vaccinations in our follow-up data collection. However, this confidence was also accompanied by elevated levels of concern toward childhood vaccines among American parents. We then examined how different subgroups in the population scored on these measures before and after the outbreak. We found that parents with high levels of interest in the topic of vaccines and a child who is not fully upto date with the recommended vaccination schedule reported more supportive attitudes toward vaccines. However, future intentions to follow the recommended vaccination schedule were not positively impacted by the outbreak. Possible explanations for these results and implications for vaccination outreach are discussed.

Exploring The Impact Of The US Measles Outbreak On Parental Awareness Of And Support For Vaccination.

Despite consensus among health officials that childhood immunizations are a safe and effective means of protecting people from disease, some parents remain hesitant about vaccinating their children. This hesitancy has been linked to a lack of confidence in recommended vaccinations as well as a desire to delay or further space out scheduled vaccinations but also outright refusal of vaccines. Using two national surveys of parents of children ages five and younger, collected immediately prior to and in the weeks following the 2014-15 US measles outbreak, this study examined the awareness of this vaccine-preventable disease outbreak among parents and whether awareness of the outbreak affected their beliefs about childhood vaccination, confidence, and intentions. The study found that while most parents were aware of the outbreak, many were not, and the level of familiarity mattered, particularly on measures of confidence in vaccines and support for mandates requiring childhood vaccination. Increases in vaccine-related concerns were found as well, indicating that disease outbreaks foster not just awareness of vaccines and their potential to prevent disease but a range of parental responses.