Our paper "Tracking the impact of COVID-19 and lockdown policy on public mental health using social media: an infoveillance study" has been accepted by JMIR JMIR Publications We built a useful pipeline with #NLP techniques to track public mental health during the pandemic.

New in JMIR: Tracking the Impact of #COVID19 #coronavirus and Lockdown Policies on Public #MentalHealth Using #SocialMedia #hcsm #SoMe: Infoveillance #Study dlvr.it/Sb1bB8

Our paper "Exploring Social Media for Early Detection of Depression in COVID-19 Patients" was accepted by #WWW2023 The Web Conference . Tweets of > 10k COVID-19 Patients were collected. Free text and mood info. were encoded with a BERT-based NLP framework. Pdf: arxiv.org/pdf/2302.12044…

Our recent JMIR Publications paper examines #COVID19 symptoms and their co-occurrence networks through #socialmedia. Social media can complement hospital-based data in public health research, particularly in comprehending patients with mild symptoms.

Our previous work in JMIR Publications showed that LLMs face challenges on Chinese clinical notes.jmir.org/2024/1/e51926/ In our recent paper JAMIA, doi.org/10.1093/jamia/…, we demonstrated that local clinical knowledge can greatly enhance LLM capabilities in this scenario.

Our recent paper "Clinical Text Datasets for Medical Artificial Intelligence and Large Language Models — A Systematic Review" has been published in NEJM AI ! The lack of clinical text data is a long-term challenge/pain for #AIinMedicine and clinical LLM researchers. After

LLMs show promise in clinical predictions, but generating reliable prediction probabilities is challenging. Our study finds that explicit probabilities from text generation often underperform—caution is needed in clinical decisions. Preprint: arxiv.org/pdf/2408.11316 #AIinMedicine

Day 1 of the 8th ISoP Seminar in Boston! 🌟 Jim Barrett (Uppsala Monitoring Centre) shared insights on critically evaluating AI in Pharmacovigilance, while Jie Yang (Harvard Medical School) explored the power of Large Language Models in EHR understanding. 📸💡 #ISoPBostonSeminar

Check out our latest publication in npj Digital Medicine: "Probabilistic Medical Predictions of Large Language Models" 🎉 We reveal key differences in how LLMs generate prediction probabilities (or confidence) and emphasize the need for caution in their clinical use. Link:

In a study npj Digital Medicine led by Bowen Gu and Jie Yang, PhD, FAMIA, we find that LLMs struggle to convey uncertainty and can be overly confident in their answer even when it is wrong. Imp area for research and improvement for detection of hallucinations.. nature.com/articles/s4174…

Impressive new results from our BRIDGE medical benchmark! The recently released MedGemma model (27B) from Google DeepMind outperforms all open-source LLMs—including the full version of DeepSeek-R1 (671B)—under 5-shot settings, showcasing its strong capability in real-world

Multiple-choice medical exams oversimplify the complexity of medicine. Try our BRIDGE benchmark, which is based on 87 real-world clinical tasks and includes far more complex, realistic tasks. We have evaluated 95 LLMs with over 3.4 billion LLM inferences arxiv.org/abs/2504.19467