Combining Hierachical VAEs with LLMs for clinically meaningful timeline summarisation in social media
This work addresses mental health monitoring by providing clinicians with improved timeline summaries from social media, though it is incremental as it builds on existing VAE and LLM methods.
The authors tackled the problem of generating clinically meaningful summaries from social media timelines for mental health monitoring by combining a hierarchical VAE with LLMs, resulting in summaries that were more factual, logically coherent, and superior to LLM-only approaches in capturing temporal changes.
We introduce a hybrid abstractive summarisation approach combining hierarchical VAE with LLMs (LlaMA-2) to produce clinically meaningful summaries from social media user timelines, appropriate for mental health monitoring. The summaries combine two different narrative points of view: clinical insights in third person useful for a clinician are generated by feeding into an LLM specialised clinical prompts, and importantly, a temporally sensitive abstractive summary of the user's timeline in first person, generated by a novel hierarchical variational autoencoder, TH-VAE. We assess the generated summaries via automatic evaluation against expert summaries and via human evaluation with clinical experts, showing that timeline summarisation by TH-VAE results in more factual and logically coherent summaries rich in clinical utility and superior to LLM-only approaches in capturing changes over time.