Zhongshen Li

Zhongshen Li

Auto-generated profile, based on publicly available information

Zhongshen Li

City University of Hong Kong

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

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

Representative Publications

View All Publications
1 June 2026 in IEEE Transactions on Evolutionary Computation
Verified
Journal Article
Closed access
SYShen You
Shen You
KZKai Zhou
Kai Zhou
ZLZhongshen Li
Zhongshen Li
KTKay Chen Tan
Kay Chen Tan
QLQiuzhen Lin
Qiuzhen Lin
Ka-Chun WongKa-Chun Wong
Ka-Chun Wong
Despite the popularity of graph neural networks, perturbed graph data is still a serious threat towards its inherent vulnerabilities. Adversarial examples can still easily manipulate the output of graph neural networks across various attack scenarios. Meanwhile, attacks on graph networks also appear to be crucial, as it can help model designers enhance the robustness of their models. In this study, we propose a sensitivity-oriented poisoning attack for self-supervised graph embedding models through bilevel optimization, which employs different optimization methods at each level. In addition, in order to improve attack effectiveness, we analyze graph structure to identify sensitive nodes and edges that guide attack directions, combining gradient-based and query-based methods to target both edge connections and node attributes. Besides, according to the defects of existing graph masked auto-encoders models, we design the feature sensitivity and feature variance to reduce the feature differentiability, which impairs the performance of the downstream model. Ablation studies validate our operator is effective on three citation datasets. And benchmark-based experiments support the effectiveness of our method on three different graph tasks. Specifically, our approach can achieve an average reduction of 3% in the accuracy of node classification compared to existing methods for attacking neural structures alone. For attacking both graph structures and attributes, our model has even achieved an average reduction of 4.5% for the node classification task, outperforming the existing methods.

2 December 2025 in Genomics, Proteomics & Bioinformatics
Verified
Journal Article
Open access
WLWentao Long
Wentao Long
ZLZhongshen Li
Zhongshen Li
JJJunru Jin
Junru Jin
JQJianbo Qiao
Jianbo Qiao
YWYu Wang
Yu Wang
LWLeyi Wei
Leyi Wei
A small peptide encoded by a non-coding RNA (ncRNA), known as a non-coding peptide (ncPEP), is emerging as a critical regulator and biomarker in cancer, holding immense promise for immunotherapy. However, the systematic identification of ncPEPs remains a challenge because existing computational methods typically analyze peptides based on sequence alone. Sequence-only analysis overlooks the fundamental biological principle that multiple distinct peptides can be translated from a single non-coding RNA transcript, thus sharing a common transcriptional origin. Here, we address this limitation by developing HGCPep, a deep learning framework that leverages hypergraphs to model these intrinsic relationships. In our model, each ncRNA is represented as a hyperedge connecting the set of peptides it encodes, thereby enriching peptide feature representations with transcriptional context. We demonstrate that HGCPep, which integrates a hypergraph neural network with a convolutional neural network, outperforms state-of-the-art methods in identifying cancer-associated ncPEPs. Furthermore, dimensionality reduction of the learned embeddings reveals distinct clustering of ncPEPs by cancer type, illustrating how the model effectively deciphers complex biological associations. Our work introduces a new method for ncPEP analysis and provides a powerful tool for discovering novel therapeutic targets in oncology. The dataset and source code of our proposed method can be found via https://github.com/Longwt123/HGCPep_Github.

26 January 2026 in IEEE Transactions on Evolutionary Computation
Verified
Journal Article
Closed access
SYShen You
Shen You
WJWei Jiang
Wei Jiang
HMHefei Mei
Hefei Mei
DGDanei Gong
Danei Gong
ZLZhongshen Li
Zhongshen Li
Ka-Chun WongKa-Chun Wong
Ka-Chun Wong
Large Language Models (LLMs) demonstrate promising capabilities in natural language understanding and reasoning with enormous parameter spaces and vast amounts of training data. These attributes have facilitated their deployment into diverse application domains. However, the underlying parameters implicitly assume decision-making boundaries, resulting in a significant number of decision spaces not covered by training data. This makes them susceptible to adversarial manipulations through carefully crafted inputs. To illuminate the vulnerabilities of LLMs, we propose a unified token-level jailbreaking attack that makes victim models generate responses for potentially harmful queries. Specifically, we propose an evolutionary algorithm to evolve perturbation sets, utilizing gradient-based and crossover-based operators to enhance performance under multiple scenarios. Furthermore, we develop a repository for reusing past perturbations and conduct an analysis of token sensitivity within LLMs, facilitating zero-shot attacks with convergence. Extensive benchmark experiments validate the effectiveness of our method on three different models, achieving increases of 62.36%, 57.89%, and 64.81% in attack success rates compared to baseline methods under white-box scenarios. In addition, the evaluation experiments demonstrate our method is effective for multiple scenarios and different size of models. This research reveals vulnerabilities of LLMs, provides theoretical foundations for developing more robust defense strategies, and contributes to building more reliable AI systems.

7 April 2025 in Knowledge-Based Systems
Verified
Journal Article
Open access
ZLZhongshen Li
Zhongshen Li
JYJixiang Yu
Jixiang Yu
SYShen You
Shen You
LWLeyi Wei
Leyi Wei
QLQiuzhen Lin
Qiuzhen Lin
KWKa-Chun Wong
Ka-Chun Wong

3 March 2026 in IEEE Transactions on Computational Biology and Bioinformatics
Verified
Journal Article
Closed access
ZLZhongshen Li
Zhongshen Li
JYJixiang Yu
Jixiang Yu
SYShen You
Shen You
HLHao Liu
Hao Liu
LCLuyang Cai
Luyang Cai
Ka-Chun WongKa-Chun Wong
Ka-Chun Wong
Biological age provides a more direct reflection of physiological status than chronological age, serving as a vital measure to evaluate health risks and aging interventions. While steroid metabolomics offers rich information for exploring aging mechanisms, the complex and nonlinear interactions within metabolic networks remain challenging in modeling. Here, we propose and describe METRON as a deep learning framework to predict biological ages from steroid metabolomics. Specifically, a Metabolite Interaction Perception Module (MIPM) is proposed to capture the interactions. Subsequently, a Group-Rational Kolmogorov-Arnold Network is also integrated to capture intricate dependencies and enhance the representation capability. We demonstrate that METRON achieves promising performance as compared to other machine learning and deep learning methods. Beyond performance, METRON offers interpretability by recovering the established markers such as Dehydroepiandrosterone (DHEA) and identifying 17-hydroxyprogesterone (17-OH-P4) as the key signature linked to hypothalamic-pituitary-adrenal axis dynamics. These results support the capacity of METRON not only to estimate biological age but also to uncover underappreciated metabolic drivers behind aging.

17 February 2025 in bioRxiv
Verified
Preprint
Open access
PDF file
WLWentao Long
Wentao Long
ZLZhongshen Li
Zhongshen Li
JJJunru Jin
Junru Jin
JQJianbo Qiao
Jianbo Qiao
YWYu Wang
Yu Wang
LWLeyi Wei
Leyi Wei
The emergence of non-coding RNA-encoded small peptides (ncPEPs) has sparked significant interest within the realm of cancer immunotherapy, owing to their potential as valuable therapeutic targets and biomarkers. The identification and characterization of cancer-associated ncPEPs assume a crucial role in propelling cancer research forward and augmenting our comprehension of immune-related processes. However, prevailing methods for cancer-associated ncPEPs identification primarily rely on sequence order, neglecting the latent relationships between peptides. As widely acknowledged in the central dogma, peptides are translated from RNA via a many-to-many mapping. In this study, we capitalize on the translation relationship and introduce a hypergraph-based model called HGCPep to enhance the representation of peptide features. Specifically, RNA is regarded as a hyperedge connecting the peptides that are translated from it. Our experimental results validate that the HGCPep approach, leveraging the strengths of hypergraph and convolutional neural networks, outperforms alternative methodologies in the recognition of ncPEPs. Additionally, employing a reduction tool, we visualize the results of multi-label classification within a 2D latent space, shedding light on how multi-label classification task influence the representation of peptides. The dataset and source code of our proposed method can be found via https://github.com/Longwt123/HGCPep_Github .