Jingcheng Du

Jingcheng Du

Auto-generated profile, based on publicly available information

Jingcheng Du

IMO Health, Rosemont, IL, United States

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 December 2019 in Journal of the American Medical Informatics Association
Verified
Journal Article
Closed access
PDF file
SWStephen Wu
Stephen Wu
KRKirk Roberts
Kirk Roberts
SDSurabhi Datta
Surabhi Datta
JDJingcheng Du
Jingcheng Du
ZJZongcheng Ji
Zongcheng Ji
HXHua Xu
Hua Xu
Objective: This article methodically reviews the literature on deep learning (DL) for natural language processing (NLP) in the clinical domain, providing quantitative analysis to answer 3 research questions concerning methods, scope, and context of current research.Materials and Methods: We searched MEDLINE, EMBASE, Scopus, the Association for Computing Machinery Digital Library, and the Association for Computational Linguistics Anthology for articles using DL-based approaches to NLP problems in electronic health records. After screening 1,737 articles, we collected data on 25 variables across 212 papers.Results: DL in clinical NLP publications more than doubled each year, through 2018. Recurrent neural networks (60.8%) and word2vec embeddings (74.1%) were the most popular methods; the information extraction tasks of text classification, named entity recognition, and relation extraction were dominant (89.2%). However, there was a “long tail” of other methods and specific tasks. Most contributions were methodological variants or applications, but 20.8% were new methods of some kind. The earliest adopters were in the NLP community, but the medical informatics community was the most prolific.Discussion: Our analysis shows growing acceptance of deep learning as a baseline for NLP research, and of DL-based NLP in the medical community. A number of common associations were substantiated (eg, the preference of recurrent neural networks for sequence-labeling named entity recognition), while others were surprisingly nuanced (eg, the scarcity of French language clinical NLP with deep learning).Conclusion: Deep learning has not yet fully penetrated clinical NLP and is growing rapidly. This review highlighted both the popular and unique trends in this active field.

27 January 2024 in Journal of the American Medical Informatics Association
Verified
Journal Article
Open access
PDF file
YHYan Hu
Yan Hu
QCQingyu Chen
Qingyu Chen
JDJingcheng Du
Jingcheng Du
XPXueqing Peng
Xueqing Peng
VKVipina Kuttichi Keloth
Vipina Kuttichi Keloth
HXHua Xu
Hua Xu
Importance: The study highlights the potential of large language models, specifically GPT-3.5 and GPT-4, in processing complex clinical data and extracting meaningful information with minimal training data. By developing and refining prompt-based strategies, we can significantly enhance the models’ performance, making them viable tools for clinical NER tasks and possibly reducing the reliance on extensive annotated datasets. Objectives: This study quantifies the capabilities of GPT-3.5 and GPT-4 for clinical named entity recognition (NER) tasks and proposes task-specific prompts to improve their performance. Materials and Methods: We evaluated these models on 2 clinical NER tasks: (1) to extract medical problems, treatments, and tests from clinical notes in the MTSamples corpus, following the 2010 i2b2 concept extraction shared task, and (2) to identify nervous system disorder-related adverse events from safety reports in the vaccine adverse event reporting system (VAERS). To improve the GPT models' performance, we developed a clinical task-specific prompt framework that includes (1) baseline prompts with task description and format specification, (2) annotation guideline-based prompts, (3) error analysis-based instructions, and (4) annotated samples for few-shot learning. We assessed each prompt's effectiveness and compared the models to BioClinicalBERT. Results: Using baseline prompts, GPT-3.5 and GPT-4 achieved relaxed F1 scores of 0.634, 0.804 for MTSamples and 0.301, 0.593 for VAERS. Additional prompt components consistently improved model performance. When all 4 components were used, GPT-3.5 and GPT-4 achieved relaxed F1 socres of 0.794, 0.861 for MTSamples and 0.676, 0.736 for VAERS, demonstrating the effectiveness of our prompt framework. Although these results trail BioClinicalBERT (F1 of 0.901 for the MTSamples dataset and 0.802 for the VAERS), it is very promising considering few training samples are needed. Discussion: The study’s findings suggest a promising direction in leveraging LLMs for clinical NER tasks. However, while the performance of GPT models improved with task-specific prompts, there's a need for further development and refinement. LLMs like GPT-4 show potential in achieving close performance to state-of-the-art models like BioClinicalBERT, but they still require careful prompt engineering and understanding of task-specific knowledge. The study also underscores the importance of evaluation schemas that accurately reflect the capabilities and performance of LLMs in clinical settings. Conclusion: While direct application of GPT models to clinical NER tasks falls short of optimal performance, our task-specific prompt framework, incorporating medical knowledge and training samples, significantly enhances GPT models' feasibility for potential clinical applications.

6 April 2025 in Nature Communications
Verified
Journal Article
Open access
QCQingyu Chen
Qingyu Chen
YHYan Hu
Yan Hu
XPXueqing Peng
Xueqing Peng
QXQianqian Xie
Qianqian Xie
QJQiao Jin
Qiao Jin
HXHua Xu
Hua Xu
The rapid growth of biomedical literature poses challenges for manual knowledge curation and synthesis. Biomedical Natural Language Processing (BioNLP) automates the process. While Large Language Models (LLMs) have shown promise in general domains, their effectiveness in BioNLP tasks remains unclear due to limited benchmarks and practical guidelines. We perform a systematic evaluation of four LLMs—GPT and LLaMA representatives—on 12 BioNLP benchmarks across six applications. We compare their zero-shot, few-shot, and fine-tuning performance with the traditional fine-tuning of BERT or BART models. We examine inconsistencies, missing information, hallucinations, and perform cost analysis. Here, we show that traditional fine-tuning outperforms zero- or few-shot LLMs in most tasks. However, closed-source LLMs like GPT-4 excel in reasoning-related tasks such as medical question answering. Open-source LLMs still require fine-tuning to close performance gaps. We find issues like missing information and hallucinations in LLM outputs. These results offer practical insights for applying LLMs in BioNLP.

4 February 2019 in BMC Genomics
Verified
Journal Article
Open access
PDF file
JDJingcheng Du
Jingcheng Du
PJPeilin Jia
Peilin Jia
YDYulin Dai
Yulin Dai
CTCui Tao
Cui Tao
ZZZhongming Zhao
Zhongming Zhao
DZDegui Zhi
Degui Zhi
Existing functional description of genes are categorical, discrete, and mostly through manual process. In this work, we explore the idea of gene embedding, distributed representation of genes, in the spirit of word embedding. From a pure data-driven fashion, we trained a 200-dimension vector representation of all human genes, using gene co-expression patterns in 984 data sets from the GEO databases. These vectors capture functional relatedness of genes in terms of recovering known pathways - the average inner product (similarity) of genes within a pathway is 1.52X greater than that of random genes. Using t-SNE, we produced a gene co-expression map that shows local concentrations of tissue specific genes. We also illustrated the usefulness of the embedded gene vectors, laden with rich information on gene co-expression patterns, in tasks such as gene-gene interaction prediction. We proposed a machine learning method that utilizes transcriptome-wide gene co-expression to generate a distributed representation of genes. We further demonstrated the utility of our distribution by predicting gene-gene interaction based solely on gene names. The distributed representation of genes could be useful for more bioinformatics applications.

23 July 2018 in BMC Medical Informatics and Decision Making
Verified
Journal Article
Open access
PDF file
JDJingcheng Du
Jingcheng Du
YZYaoyun Zhang
Yaoyun Zhang
JLJianhong Luo
Jianhong Luo
YJYuxi Jia
Yuxi Jia
QWQiang Wei
Qiang Wei
HXHua Xu
Hua Xu
Suicide has been one of the leading causes of deaths in the United States. One major cause of suicide is psychiatric stressors. The detection of psychiatric stressors in an at risk population will facilitate the early prevention of suicidal behaviors and suicide. In recent years, the widespread popularity and real-time information sharing flow of social media allow potential early intervention in a large-scale population. However, few automated approaches have been proposed to extract psychiatric stressors from Twitter. The goal of this study was to investigate techniques for recognizing suicide related psychiatric stressors from Twitter using deep learning based methods and transfer learning strategy which leverages an existing annotation dataset from clinical text. First, a dataset of suicide-related tweets was collected from Twitter streaming data with a multiple-step pipeline including keyword-based retrieving, filtering and further refining using an automated binary classifier. Specifically, a convolutional neural networks (CNN) based algorithm was used to build the binary classifier. Next, psychiatric stressors were annotated in the suicide-related tweets. The stressor recognition problem is conceptualized as a typical named entity recognition (NER) task and tackled using recurrent neural networks (RNN) based methods. Moreover, to reduce the annotation cost and improve the performance, transfer learning strategy was adopted by leveraging existing annotation from clinical text. To our best knowledge, this is the first effort to extract psychiatric stressors from Twitter data using deep learning based approaches. Comparison to traditional machine learning algorithms shows the superiority of deep learning based approaches. CNN is leading the performance at identifying suicide-related tweets with a precision of 78% and an F-1 measure of 83%, outperforming Support Vector Machine (SVM), Extra Trees (ET), etc. RNN based psychiatric stressors recognition obtains the best F-1 measure of 53.25% by exact match and 67.94% by inexact match, outperforming Conditional Random Fields (CRF). Moreover, transfer learning from clinical notes for the Twitter corpus outperforms the training with Twitter corpus only with an F-1 measure of 54.9% by exact match. The results indicate the advantages of deep learning based methods for the automated stressors recognition from social media.

24 June 2019 in Journal of the American Medical Informatics Association
Verified
Journal Article
Open access
PDF file
JDJingcheng Du
Jingcheng Du
QCQingyu Chen
Qingyu Chen
YPYifan Peng
Yifan Peng
YXYang Xiang
Yang Xiang
CTCui Tao
Cui Tao
ZLZhiyong Lu
Zhiyong Lu
In multi-label text classification, each textual document is assigned 1 or more labels. As an important task that has broad applications in biomedicine, a number of different computational methods have been proposed. Many of these methods, however, have only modest accuracy or efficiency and limited success in practical use. We propose ML-Net, a novel end-to-end deep learning framework, for multi-label classification of biomedical texts. ML-Net combines a label prediction network with an automated label count prediction mechanism to provide an optimal set of labels. This is accomplished by leveraging both the predicted confidence score of each label and the deep contextual information (modeled by ELMo) in the target document. We evaluate ML-Net on 3 independent corpora in 2 text genres: biomedical literature and clinical notes. For evaluation, we use example-based measures, such as precision, recall, and the F measure. We also compare ML-Net with several competitive machine learning and deep learning baseline models. Our benchmarking results show that ML-Net compares favorably to state-of-the-art methods in multi-label classification of biomedical text. ML-Net is also shown to be robust when evaluated on different text genres in biomedicine. ML-Net is able to accuractely represent biomedical document context and dynamically estimate the label count in a more systematic and accurate manner. Unlike traditional machine learning methods, ML-Net does not require human effort for feature engineering and is a highly efficient and scalable approach to tasks with a large set of labels, so there is no need to build individual classifiers for each separate label.