Adverse drug events (ADEs), referring to harmful and unintended reactions from routine medication use, continue to pose a significant challenge for pharmacovigilance, one that can be addressed through AI-enabled Web services. In the United States, more than two million serious ADEs have been reported among hospitalized patients. These instances can be reduced if hospitals and health centers use systems to identify, anticipate, or estimate the likelihood of ADE occurrences documented on clinical notes. While computational methods leveraging natural language processing (NLP) have been developed for ADE detection in clinical texts, most of them rely on classical machine learning or deep learning models that often fail to incorporate domain-specific biomedical knowledge. We propose a knowledge graph-augmented large language model (LLM) framework for ADE prediction. Our approach uses zero shot prompting and instruction-based prompting for finetuning, while integrating structured knowledge from a Drug Adverse Event Knowledge Graph (KG) constructed with the Ontology of Adverse Events (OAE), RxNorm, and the Cadec ADE dataset. Experiments on ADE prediction indicate that KG-enhanced LLMs outperform standard LLM-based approaches, achieving up to an 8% improvement in F1 score with Mistral 7B and a 16% boost with GPT-4.1 on instruction-tuned ADE data, as well as a 1% and 2% F1-score increase, respectively, on instruction-tuned TAC data. These results suggest the effectiveness of combining symbolic biomedical knowledge with LLMs for advancing ADE prediction and improving drug safety in clinical settings.
Shirish Bajpai, V. S. S. Anirudh Sharma, Aman Tomar et al.· International Conference on...· 0 citations
Deep learning models in computer vision face significant challenges when trained on long-tailed datasets, where a few majority classes dominate while many minority classes are severely underrepresented. Such imbalances frequently arise in real-world scenarios such as rare species recognition, manufacturing fault detection, and medical image understanding, leading to biased models that underperform on tail classes. Existing reweighting methods typically rely on static class frequencies to penalize the model, ignoring the dynamic nature of how effectively a network actually learns a class over time. We address this by introducing a novel Learning-Dynamics Aware Loss (LDAL) function that shifts the focus from static sample counts to dynamic learning progress. LDAL framework adjusts class weights continuously by leveraging: (i) the strength of learned feature representations (semantic scale), (ii) the intrinsic learning difficulty of each class, measured via the Shannon entropy of its predictions, and (iii) an inter-epoch regularizer term that tracks prediction shifts between consecutive epochs to stabilize training and avoid local minima. LDAL is purely a objective function which incurs negligible computational overhead while adapting to the feature learning of the model. Experimental results on multiple benchmark datasets demonstrate that our approach significantly surpasses state-of-the-art reweighting loss functions, providing an optimal trade-off between accuracy and generalizability. The source code is available at https://github.com/sdm2026/ldal
V. Shinde, Nikhil Kumar Shrey, Magesh Rajasekaran et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
