Problem being addressed
In a clinical decision support system, doctors and healthcare professionals require access to information from heterogeneous sources, such as research papers, electronic health records clinical case reports, reference works and knowledge base articles; however, they still lack efficient answer retrieval from long healthcare documents.
A query model that is focused on clinical aspects such as therapy, diagnosis, etiology, prognosis, and others, which have been described in the literature previously by manual clustering of semantic question types or crawling medical Wikipedia section headings. The model uses multi-task learning to align the sequence of sentences in a long document, which allows to answer ad-hoc queries with short latency.
Advantages of this solution
The generalized model signifficantly outperforms several state-of-the-art baselines for healthcare passage ranking and is able to adapt to heterogeneous domains without additional finetuning. It is the first method to address answer retrieval with structured queries on long heterogeneous documents from the healthcare domain.
Solution originally applied in these industries
Possible New Application of the Work
The suggested approach can be incorporated in more general and not industry-specific business intelligence systems. This will allow large enterprises better manage their non-homogenous data sources and obtain more accurate query results.
Source URL: #############