Hailong Jiang

Hailong Jiang

Auto-generated profile, based on publicly available information

Hailong Jiang

Department of Computer Science, Kent State University

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

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

Representative Publications

View All Publications
5 March 2020 in IEEE Access
Verified
Journal Article
Open access
ALAo Li
Ao Li
YDYu Ding
Yu Ding
DCDeyun Chen
Deyun Chen
GSGuanglu Sun
Guanglu Sun
HJHailong Jiang
Hailong Jiang
QWQidi Wu
Qidi Wu
The cross-view multimedia are widely existed and attract many attentions in recent years. Nevertheless, it is noted that the phenomenon, that data in different classes from same view are more similar than that in same class from different views, is usually presented for cross-view multimedia data. The intrinsic imperfection leads disappointing performance for cross-view multimedia recognition or classification. To solve this problem, in this paper, we propose a novel discriminative learning framework with low-rank constraint, which can be applied for view-invariant low-dimensional subspace learning. The advantages of our framework include three aspects. Firstly, to unlock the latent class structure and view structure, a self-expressed model by dual low-rank constraints are presented, which can separate the two manifold structures in the learned subspace. Secondly, two effective discriminative graphs are constructed to guide the affinity relationship of data in the above two low-dimensional projected subspaces respectively. Finally, the joint semantic consensus constraint is designed to be integrated into the learning framework, which can explore the shared and view-specific information for enforcing the view-invariant character in semantic space. Experimental results on several public cross-view multimedia datasets demonstrate that our proposed method outperforms existing excellent subspace learning approaches.

26 February 2025 in JMIR Formative Research
Verified
Journal Article
Open access
JZJianfeng Zhu
Jianfeng Zhu
XZXinyu Zhang
Xinyu Zhang
RJRuoming Jin
Ruoming Jin
HJHailong Jiang
Hailong Jiang
DKDeric R Kenne
Deric R Kenne
Background: Antidepressants are crucial for managing major depressive disorders; however, nonadherence remains a widespread challenge, driven by concerns over side effects, fear of dependency, and doubts about efficacy. Understanding patients’ experiences is essential for improving patient-centered care and enhancing adherence, which prioritizes individual needs in treatment. Objective: This study aims to gain a deeper understanding of patient experiences with antidepressants, providing insights that health care providers, families, and communities can develop into personalized treatment strategies. By integrating patient-centered care, these processes may improve satisfaction and adherence with antidepressants. Methods: Data were collected from AskaPatient and Reddit, analyzed using natural language processing and large language models. Analytical techniques included sentiment analysis, emotion detection, personality profiling, and topic modeling. Furthermore, demographic variations in patient experiences were also examined to offer a comprehensive understanding of discussions around antidepressants. Results: Sentiment and emotion analysis revealed that the majority of discussions (21,499/36,253, 59.3%) expressed neutral sentiments, with negative sentiments following closely (13,922/36,253, 38.4%). The most common emotions were fear (16,196/36,253, 44.66%) and sadness (12,507/36,253, 34.49%). The largest topic, “Mental Health and Relationships,” accounted for 11.69% (3755/36,253) of the discussions, which indicated a significant focus on managing mental health conditions. Discussions around nonadherence were marked by fear, followed by sadness, while self-care discussions showed a notable trend of sadness. Conclusions: These psychological insights into public perceptions of antidepressants provide a foundation for developing tailored, patient-centered treatment approaches that align with individual needs, enhancing both effectiveness and empathy of care.

25 August 2026 in Applied Intelligence
Verified
Journal Article
Closed access
PDF file
GZGuilin Zhang
Guilin Zhang
WGWulan Guo
Wulan Guo
ZTZiqi Tan
Ziqi Tan
PWPengcheng Wang
Pengcheng Wang
HJHailong Jiang
Hailong Jiang

28 May 2024 in JMIR Preprints
Verified
Preprint
Open access
JZJianfeng Zhu
Jianfeng Zhu
XZXinyu Zhang
Xinyu Zhang
RJRuoming Jin
Ruoming Jin
HJHailong Jiang
Hailong Jiang
DKDeric R Kenne
Deric R Kenne
BACKGROUND Antidepressant medications are widely used to manage depression and anxiety disorders. Given that these medications have been around for many years, their side effects are well documented. Apart from common side effects, there are new emerging data reporting unexpected side effects or stress related challenges that clinicians should be aware of. OBJECTIVE The goal of this study to analyze the demographic profiles, symptomatology, and platform-specific differences in online discussions about antidepressant side effects. Exploring the antidepressant side effects in different gender and age, then the clinical with detailed information to improve the documentation and management of drug side effects. METHODS We collected dataset from Ask a patient and Reddit platform. A mixed methods approach including quantitative analysis and qualitative analysis. Which we leverage Natural Language Processing method to examine the demographic and symptoms comments, and exploring the language linguistic patterns on word, topic, and theme of posts. RESULTS The analysis highlighted significant gender-based differences in antidepressant side effects. Females reported more gastrointestinal, neurological, sleep, general physical, and emotional symptoms, while males reported more sexual health issues. Users frequently discussed withdrawal symptoms and challenges with switching medications, revealing anxiety about the withdrawal process. Unique symptoms such as toe infections, visual changes, and pulsatile tinnitus were also noted. The strong correlation between fear and neurological, gastrointestinal, and sleep-related symptoms suggests these physical symptoms significantly contribute to emotional distress. CONCLUSIONS These findings underscore the necessity for gender-sensitive healthcare approaches and highlight new concerns that are not typically covered in standard drug descriptions. Addressing these issues directly in consultations can improve patient outcomes by fostering better adherence to medication regimens and enhancing overall treatment satisfaction. Future research should focus on understanding the unique challenges faced by specific demographics, particularly women in reproductive age who use birth control, to provide more effective and personalized care.

15 June 2025 in TechRxiv
Verified
Preprint
Open access
HJHailong Jiang
Hailong Jiang
JZJianfeng Zhu
Jianfeng Zhu
BFBo Fang
Bo Fang
KBKevin Barker
Kevin Barker
CCChao Chen
Chao Chen
QGQiang Guan
Qiang Guan
Predicting the soft error resilience of highperformance computing (HPC) applications is critical for deploying reliable systems, yet traditional fault injection (FI) approaches are costly and difficult to scale. We present eHAPPA, a lightweight, parameter-efficient framework that estimates resilience scores directly from source code using pretrained language models. eHAPPA builds on CodeBERT and introduces three key innovations: (1) a chunk-based encoding pipeline that handles long code sequences via tokenization, overlapping segmentation, and contextual embedding; (2) a comparative evaluation of aggregation strategies-including mean, max, LSTM, and attention pooling-to integrate chunk-level representations; and (3) the integration of Low-Rank Adaptation (LoRA), which enables scalable fine-tuning with less than 1% of parameters updated. We conduct extensive experiments across multiple configurations of LoRA rank, chunk overlap rate, and pooling method. Results show that eHAPPA consistently outperforms prior approaches such as HAPPA, achieving up to 24.6% reduction in MSE while significantly reducing computational overhead. Attention pooling combined with moderate LoRA ranks (e.g., 16) yields the best performance, demonstrating the value of expressive context modeling even under small-data regimes. Our findings suggest that pretrained models, when adapted with efficient tuning and principled aggregation, offer a powerful alternative to traditional FI-based resilience evaluation. eHAPPA provides a modular foundation for static, low-cost, and high-quality resilience estimation, with potential extensions to broader program analysis tasks such as vulnerability detection and compiler-guided optimization.