In an era where every public health action is expected to be backed by credible evidence, health policy-making has also been increasingly seen to follow the same. The general consensus across the globe is to strengthen health information systems and the decisions of the policy makers are increasingly relying on the information provided to them through such systems. COVID-19 has clearly brought out the need for accurate, timely and relevant information in planning for and responding to public health emergencies that can be equally devastating, if not more. It is crucial for information providers to understand the importance of communicating and disseminating it in a timely manner so that it leads to public health action for the larger good of the population.

Automatically generating a presentation from the text of a long document is a challenging and useful problem. In contrast to a flat summary, a presentation needs to have a better and non-linear narrative, i.e., the content of a slide can come from different and non-contiguous parts of the given document. However, it is difficult to incorporate such non-linear mapping of content to slides and ensure that the content is faithful to the document. LLMs are prone to hallucination and their performance degrades with the length of the input document. Towards this, we propose a novel graph based solution where we learn a graph from the input document and use a combination of graph neural network and LLM to generate a presentation with attribution of content for each slide. We conduct thorough experiments to show the merit of our approach compared to directly using LLMs for this task.

No abstract