A population and public health NLP solution from H2O.ai Health
Powered by NVIDIA GPUs and NVIDIA AI
Social media platforms such as Twitter and Reddit have become invaluable tools for communication between individuals or groups and are widely used globally. As messages on these platforms can instantly be accessed by all users and remain on communication threads until deleted, specific messages have been and will continue to influence individual opinions towards many topics, including vaccination for a considerable amount of time.
Unfortunately, these platforms have fostered a rise in the spread of unsubstantiated and false information regarding vaccination, leading to reduced uptake for vaccines in individuals or for their children. In addition, vaccine misinformation shared on these platforms has contributed to an increase in vaccine hesitancy and, in certain instances, vaccine preventable disease outbreaks in unvaccinated populations. Actively monitoring conversations around vaccination sentiment on social media platforms and swiftly addressing misinformation is critical to prevent translation into vaccine hesitancy or disease outbreaks.
Several studies have been conducted to analyze vaccine-related social media messages. They found that most vaccine-related tweets expressed confidence in vaccines, their safety, and the system that delivers them, while a very small minority (<5%) expressed a lack of confidence. While the percentage of messages expressing lack of confidence is relatively low, it has been shown that a small group can play an effective and outsized role in spreading misinformation through misleading narratives and thus hinder people’s ability to judge the true risks associated with vaccines.
We propose to build on the current study in the following ways:
- Omni-channel trending, topic, and sentiment analytics and AI
- Global applicability and comparative trends across locations
- Monitor data and sentiment over time with real-time knowledge creation
- Ensuring interpretability, explainability, and transparency of the NLP models
- Creating a dashboard to monitor (mis)information across channels and diverse platforms
We are continuing to work on implementing statistical metrics and NLP with purpose-built deep learning to track vaccine hesitancy, a project that is essential to informing citizens, public health professionals, and pharmaceutical real-world evidence teams. It also aims to contribute to medical and patient education, and external stakeholder empowerment. This work is already providing invaluable insights to internal cross-functional teams/leaders – as well as healthcare professionals and public health authorities – to mitigate misinformation on vaccines and to improve our global public health campaigns.
Vaccine NLP is powered by the H2O AI Cloud : Driverless AI AutoML , H2O-3, MLOps, and H2O.ai Wave©, as well as by NVIDIA GPUs.
