An official website of the United States government
Data imbalance is a fundamental challenge in ap- plying language models to biomedical applications, particularly in ICD code prediction tasks where label and demographic distributions are uneven. While state-of-the-art language models have been increasingly adopted in biomedical tasks, few studies have systematically examined how data imbalance affects model performance and fairness across demographic groups. This study fills the gap by statistically probing the relationship between data imbalance and model performance in ICD code prediction. We analyze imbalances in a standard benchmark data across gender, age, ethnicity, and social determinants of health by state- of-the-art biomedical language models. By deploying diverse performance metrics and statistical analyses, we explore the influence of data imbalance on performance variations and demographic fairness. Our study shows that data imbalance significantly impacts model performance and fairness, but feature similarity to the majority class may be a more critical factor. We believe this study provides valuable insights for developing more equitable and robust language models in healthcare applications.
more »
« less
This content will become publicly available on November 20, 2025
Time Matters: Examine Temporal Effects on Biomedical Language Models
Time roots in applying language models for biomedical applications: models are trained on historical data and will be deployed for new or future data, which may vary from training data. While increasing biomedical tasks have employed state-of-the-art language models, there are very few studies have examined temporal effects on biomedical models when data usually shifts across development and deployment. This study fills the gap by statistically probing relations between language model performance and data shifts across three biomedical tasks. We deploy diverse metrics to evaluate model performance, distance methods to measure data drifts, and statistical methods to quantify temporal effects on biomedical language models. Our study shows that time matters for deploying biomedical language models, while the degree of performance degradation varies by biomedical tasks and statistical quantification approaches. We believe this study can establish a solid benchmark to evaluate and assess temporal effects on deploying biomedical language models.
more »
« less
- Award ID(s):
- 2245920
- PAR ID:
- 10529033
- Publisher / Repository:
- AMIA Annual Symposium Proceedings
- Date Published:
- Page Range / eLocation ID:
- 723-732
- Format(s):
- Medium: X
- Location:
- San Francisco, CA
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Finetuned large language models (LLMs) have shown remarkable performance in financial tasks, such as sentiment analysis and information retrieval. Due to privacy concerns, finetuning and deploying financial LLMs (FinLLMs) locally are crucial for institutions and individuals. In this paper, we employ quantized low-rank adaptation (QLoRA) to finetune FinLLMs, which leverage low-rank structure and quantization technique to significantly reduce computational requirements while maintaining model performance. We also employ data and pipeline parallelism to enable local finetuning on commodity GPUs. Experiments on financial datasets validate the efficacy of our approach in yielding notable improvements over the base models.more » « less
-
State-space models (SSMs), such as Mamba (Gu & Dao, 2023), have been proposed as alternatives to Transformer networks in language modeling, incorporating gating, convolutions, and input-dependent token selection to mitigate the quadratic cost of multi-head attention. Although SSMs exhibit competitive performance, their in-context learning (ICL) capabilities, a remarkable emergent property of modern language models that enables task execution without parameter optimization, remain less explored compared to Transformers. In this study, we evaluate the ICL performance of SSMs, focusing on Mamba, against Transformer models across various tasks. Our results show that SSMs perform comparably to Transformers in standard regression ICL tasks, while outperforming them in tasks like sparse parity learning. However, SSMs fall short in tasks involving non-standard retrieval functionality. To address these limitations, we introduce a hybrid model, MambaFormer, that combines Mamba with attention blocks, surpassing individual models in tasks where they struggle independently. Our findings suggest that hybrid architectures offer promising avenues for enhancing ICL in language models.more » « less
-
Background: Outcome measures that are count variables with excessive zeros are common in health behaviors research. Examples include the number of standard drinks consumed or alcohol‐related problems experienced over time. There is a lack of empirical data about the relative performance of prevailing statistical models for assessing the efficacy of interventions when outcomes are zero‐inflated, particularly compared with recently developed marginalized count regression approaches for such data.Methods: The current simulation study examined five commonly used approaches for analyzing count outcomes, including two linear models (with outcomes on raw and log‐transformed scales, respectively) and three prevailing count distribution‐based models (ie, Poisson, negative binomial, and zero‐inflated Poisson (ZIP) models). We also considered the marginalized zero‐inflated Poisson (MZIP) model, a novel alternative that estimates the overall effects on the population mean while adjusting for zero‐inflation. Motivated by alcohol misuse prevention trials, extensive simulations were conducted to evaluate and compare the statistical power and Type I error rate of the statistical models and approaches across data conditions that varied in sample size ( to 500), zero rate (0.2 to 0.8), and intervention effect sizes.Results: Under zero‐inflation, the Poisson model failed to control the Type I error rate, resulting in higher than expected false positive results. When the intervention effects on the zero (vs. non‐zero) and count parts were in the same direction, the MZIP model had the highest statistical power, followed by the linear model with outcomes on the raw scale, negative binomial model, and ZIP model. The performance of the linear model with a log‐transformed outcome variable was unsatisfactory.Conclusions: The MZIP model demonstrated better statistical properties in detecting true intervention effects and controlling false positive results for zero‐inflated count outcomes. This MZIP model may serve as an appealing analytical approach to evaluating overall intervention effects in studies with count outcomes marked by excessive zeros.more » « less
-
null (Ed.)Text classification is a fundamental problem, and recently, deep neural networks (DNN) have shown promising results in many natural language tasks. However, their human-level performance relies on high-quality annotations, which are time-consuming and expensive to collect. As we move towards large inexpensive datasets, the inherent label noise degrades the generalization of DNN. While most machine learning literature focuses on building complex networks to handle noise, in this work, we evaluate model-agnostic methods to handle inherent noise in large scale text classification that can be easily incorporated into existing machine learning workflows with minimal interruption. Specifically, we conduct a point-by-point comparative study between several noise-robust methods on three datasets encompassing three popular classification models. To our knowledge, this is the first time such a comprehensive study in text classification encircling popular models and model-agnostic loss methods has been conducted. In this study, we describe our learning and demonstrate the application of our approach, which outperformed baselines by up to 10% in classification accuracy while requiring no network modifications.more » « less
