Improved Methods to Aid Unsupervised Evidence-Based Fact Checking for Online Heath News

Abstract

False information in the domain of online health related articles is of great concern, which has been witnessed abundantly in the current pandemic situation of Covid-19. Recent advancements in the field of Machine Learning and Natural Language Processing can be leveraged to aid people in distinguishing false information from the truth in the domain of online health articles. Whilst there has been substantial progress in this space over the years, research in this area has mainly focused on the sphere of political news. Health fake news is markedly different from fake news in the political context as health information should be evaluated against the most recent and reliable medical resources such as scholarly repositories. However, one of the challenges with such an approach is the retrieval of the pertinent resources. In this work, we formulate two techniques for the retrieval of the most relevant authoritative and reliable medical content from scholarly repositories which can be used to assess veracity of an online health article. The first technique is an unsupervised method of generating queries from claims which are extracted from an online health article. We propose a three-step approach for it and illustrate that our method is able to generate effective queries which can be used for retrieval of information from medical knowledge databases. The second method involves a filtering approach for extracting the most relevant information for the claims. We show how this can be achieved with the help of state of the art transformer models and illustrate it's effectiveness over other methods.

Tracing topics and trends in drug‐resistant epilepsy research using a natural language processing–based topic modeling approach

“…• Our topic modeling analysis uncovered 18 distinct themes in the drug-resistant epilepsy literature, spanning diverse areas and demonstrating the heterogeneous nature of this research field. • The 10 most common topics identified were "AEDs," "Neuromodulation Therapy," "Genomics," "Surgical Outcomes," "Source Localization," "Ketogenic Diet," "Autoimmune Encephalitis," "Disconnection Surgery," "Functional Brain Mapping," and "Focal Cortical Dysplasia."…”

“…To derive sentence embeddings, we used the “S‐PubMedBert‐MS‐MARCO” model, 18 a Hugging Face sentence‐transformers model that is fine‐tuned specifically for medical text information retrieval. As part of preprocessing, we removed frequent words or stop words that appear commonly in the text but do not provide insight into a document's specific topic, such as “the”, “and”, “of”, and so on, after the embeddings were obtained.…”

From Text to Insight: A Natural Language Processing-Based Analysis of Topics and Trends in Neurosurgery

“…To obtain sentence embeddings, we used the S-PubMedBert-MS-MARCO model, 8 a Hugging Face sentence-transformers model specifically fine-tuned for information retrieval tasks in the medical and health text domain. Sentence embeddings are vector representations of sentences, where each vector encapsulates the semantic information of the corresponding sentence in a high-dimensional space.…”

Navigating the Landscape of Precision Education: Insights From On-the-Ground Initiatives

“…The initial content items and mappings were embedded using a natural language processing (NLP) modelprtiamdeka/s-PubMedBert-MS-MARCO -an open-source model trained on articles in PubMed. 29 A weighted rules-based prioritization algorithm utilizes content already viewed by the learner, prior learner experience with procedures or medical diagnoses, and their current level of competency to prioritize the list. For example, if the learner encounters a patient with diabetes and this triggers the system to recommend an article on the intraoperative management of patients with diabetes, the prioritization algorithm would downgrade the display of this content if the learner already had a significant number of similar encounters.…”

“…The content database can be updated with new content, an approach that aligns with the principle of proactivity. The initial content items and mappings were embedded using a natural language processing (NLP) model—prtiamdeka/s-PubMedBert-MS-MARCO—an open-source model trained on articles in PubMed 29 . A weighted rules-based prioritization algorithm utilizes content already viewed by the learner, prior learner experience with procedures or medical diagnoses, and their current level of competency to prioritize the list.…”