About

Susu Zhang is an Associate Professor in the Department of Statistics at the Illinois College of Liberal Arts & Sciences. Her research focuses on how measurement theory and psychometrics can guide the assessment of both human traits, such as psychological characteristics, knowledge, and skills, and machine capabilities, including AI models. She develops new statistical tools to address practical questions in these domains that cannot be adequately handled with existing methods. Her current research interests include developing measurement theory and psychometric tools for unstructured test response data, adapting large language models to support measurement theory-grounded assessments for learning, and creating measurement theory and psychometric tools for AI model evaluation and benchmark design.

Research topics

• Biology
• Endocrinology
• Pathology
• Cell biology
• Medicine
• Internal medicine

Selected publications

• HIP: Model-agnostic hypergraph influence prediction via distance-centrality fusion and neural ODEs

  Expert Systems with Applications · 2026-04-07

  preprintOpen access1st authorCorresponding
  PublisherOA PDFDOI

• BCIM: Budget and capacity constrained influence maximization in multilayer networks

  Tsinghua Science & Technology · 2026-04-01

  preprintOpen access1st authorCorresponding

  Influence maximization (IM) seeks to identify a seed set that maximizes influence within a network, with applications in areas such as viral marketing, disease control, and political campaigns. The budgeted influence maximization (BIM) problem extends IM by incorporating cost constraints for different nodes. However, the current BIM problem, limited by budget alone, often results in the selection of numerous low-cost nodes, which may not be applicable to real-world scenarios. Moreover, considering that users can transmit information across multiple social platforms, solving the BIM problem across these platforms could lead to more optimized resource utilization. To address these challenges, we propose the Budget and Capacity Constrained Influence Maximization (BCIM) problem within multilayer networks and introduce a Multilayer Multi-population Genetic Algorithm (MMGA) to solve it. The MMGA employs modules, such as initialization, repair, and parallel evolution, designed not only to meet budget and capacity constraints but also to significantly enhance algorithmic efficiency. Extensive experiments on both synthetic and empirical multilayer networks demonstrate that MMGA improves spreading performance by at least 10% under the two constraints compared to baselines extended from classical IM problems. The BCIM framework introduces a novel direction in influence maximization, providing an effective and efficient solution to the problem.

  PublisherOA PDFDOI

• Triboelectric Separation Technology for Sustainable Recycling of Decommissioned Wind Turbine Blades

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Jointly modeling responses and omitted items by a competing risk model: A survival analysis approach

  British Journal of Mathematical and Statistical Psychology · 2025-01-30 · 1 citations

  articleOpen accessSenior authorCorresponding

  Item response theory models are commonly adopted in educational assessment and psychological measurement. Such models need to be modified to accommodate practical situations when statistical sampling assumptions are violated. Omission is a common phenomenon in educational testing. In modern computer-based testing, we have not only examinees' responses but also their response times. This paper utilizes response time and develops a joint model of responses and response times. The new approach is analogous to those developed in survival analysis for dealing with right-censored data. In particular, a key ingredient is the introduction of the omission time (OT), which corresponds to the censoring time in survival analysis. By competing risk formulation, the proposed method provides an alternative narrative to how an item becomes answered versus omitted, depending on the competing relationship of response time and OT, so that the likelihood function can be constructed properly. The maximum likelihood estimator can be computed via the expectation-maximization algorithm. Simulation studies were conducted to evaluate the performance of the proposed method and its robustness against various mis-specifications. The method was applied to a dataset from the PISA 2015 Science Test.

  PublisherOA PDFDOI

• Explaining Performance Gaps with Problem-Solving Process Data via Latent Class Mediation Analysis

  Psychometrika · 2025-08-11 · 1 citations

  articleOpen accessSenior authorCorresponding

  Process data, in particular, log data collected from a computerized test, documents the sequence of actions performed by an examinee in pursuit of solving a problem, affording an opportunity to understand test-taking behavioral patterns that account for demographic group differences in key outcomes of interest, for instance, final score on a cognitive item. Addressing this aim, this article proposes a latent class mediation analysis procedure. Using continuous process features extracted from action sequence data as indicators, latent classes underlying the test-taking behavior are identified in a latent class mediation model, where an examinee's nominal latent class membership enters as the mediator between the observed grouping and outcome variables. A headlong search algorithm for selecting the subset of process features that maximizes the total indirect effect of the latent class mediator is implemented. The proposed procedure is validated with a series of simulations. An application to a large-scale assessment highlights how the proposed method can be used to explain performance gaps between students with learning disability and their typically developing peers on the National Assessment of Educational Progress (NAEP) math assessment.

  PublisherOA PDFDOI

• CDK8 Inhibition Releases the Muscle Differentiation Block in Fusion-driven Alveolar Rhabdomyosarcoma

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-07-18

  preprintOpen access1st authorCorresponding

  ABSTRACT Alveolar rhabdomyosarcoma (aRMS) is a fusion-driven pediatric cancer with poor survival and limited therapeutic options. To uncover novel vulnerabilities, we employed complex-based analysis of the DepMap functional genomic data, identifying CDK8 as a dependency in aRMS. Both CDK8 knockout and pharmacologic inhibition impaired tumor cell growth and induced myogenic differentiation in vitro and in vivo . Compared to genetic loss, CDK8 inhibition induced more dynamic transcriptional changes. With a genome-scale CRISPR-Cas9 drug modifier screen, we determined that the maximal anti-tumor activity of the CDK8 inhibitor requires the presence of the Mediator kinase module and transcriptional cooperation with the SAGA complex. We further identified SIX4 as a key transcription factor mediating CDK8 inhibitor-induced transcriptional activation of myogenic differentiation genes and tumor cell proliferation. These findings suggest a distinct gain-of-function mechanism of the CDK8 inhibitor and establish a strong rationale for CDK8 inhibition as a differentiation-inducing therapeutic strategy in aRMS. STATEMENT OF SIGNIFICANCE We provide a framework for uncovering therapeutic targets by network-based analysis of functional genomic screens. We identify CDK8 as a druggable target in aRMS and determine that CDK8 inhibition drives myogenic differentiation and impairs tumor progression via a collaborative mechanism involving the Mediator kinase module, SAGA complex, and SIX4.

  PublisherDOI

• The soft power of kites: Leveraging the Yangjiabu Kite Festival as a model for international cultural diplomacy

  Journal of Infrastructure Policy and Development · 2025-01-10

  articleOpen access1st authorCorresponding

  The Yangjiabu Kite Festival, originating over 2000 years ago in Shandong Province, China, stands as a testament to the enduring cultural heritage and artistic traditions of kite flying. This research explores the historical origins, cultural symbolism, festival format, community engagement, and international exposure of the Yangjiabu Kite Festival, shedding light on its evolution and impact over time. Findings reveal the festival's deep roots in ancient Chinese traditions, its role as a platform for showcasing cultural diversity and craftsmanship, and its significance in promoting tourism, cultural exchange, and soft power projection for Shandong Province. Lessons learned from the Yangjiabu Kite Festival offer valuable insights for cross-cultural application, event management, cultural diplomacy, and community development. Suggestions for future research include comparative studies, longitudinal assessments, audience research, and policy analysis to further explore the dynamics and implications of cultural festivals in a global context. Overall, the research underscores the importance of cultural festivals as vehicles for cultural preservation, tourism promotion, and intercultural dialogue, fostering mutual understanding and appreciation across borders.

  PublisherOA PDFDOI

• Research on Recovery of Glass Fibers From Retired Wind Turbine Blades by Flotation and Low-Temperature Plasma Treatment

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access1st authorCorresponding
  PublisherDOI

• Identification of novel inhibitors of dengue viral NS5 RNA-dependent RNA polymerase through molecular docking, biological activity evaluation and molecular dynamics simulations

  Journal of Enzyme Inhibition and Medicinal Chemistry · 2025-02-12 · 8 citations

  articleOpen access

  =19.67 ± 1.12 μM). The comprehensive MD simulations were performed on the candidates to assess the stability behaviour and binding mechanisms. The density functional theory (DFT) analysis was also conducted to explore the structural and electronic properties.

  PublisherDOI

• Analysis of the Abnormal Characteristics of Alveolar Sounds in Postoperative Palatoplasty Patients and Rehabilitation Methods

  Journal of Craniofacial Surgery · 2025-03-31

  article

  The aim of this study was to analyze and compare the error types of alveolar sounds /s/, /ts'/, and /ts/ (ie, fricative and affricate sounds) in different ethnic groups (Han, Uyghur, Kazakh) after palatoplasty. The goal was to provide a basis for clinical assessment and effective intervention. A total of 68 patients who had undergone palatoplasty in our hospital's maxillofacial surgery department were retrospectively selected; all presented with alveolar sound abnormalities. These patients were divided into 3 age groups: 3.5 to 6 years, 7 to 13 years, and 14 years and above. Among them were 39 Han, 18 Uyghur, and 11 Kazakh patients. All had a good velopharyngeal function, normal speech organs, and occlusion, and no hearing or cognitive impairments (ie, no cognition disorders). Assessment employed the Chinese Standard Mandarin speech intelligibility scale, an alveolar sound (/s/, /ts'/, /ts/) clarity scale, the PRAAT speech workstation, and a soundproof booth equipped with a professional noise-shielding sound card and microphone. Three experienced speech therapists independently evaluated error rates and error types for each alveolar sound (/s/, /ts'/, /ts/) by different ethnic groups and ages, and they also recorded intervention duration. Error articulation patterns included omission (deletion), weakening, and various compensatory articulations. The highest error rate was found in /ts/, predominantly omission (67.65%, 46/68). Next, /s/ showed mainly omission and lateralization errors, whereas /ts'/ was primarily characterized by weakening. There was no significant difference ( P >0.05) in the incidence of alveolar sound articulation errors among different ethnic groups. Spectrogram analysis objectively confirmed whether the error type involved omission, weakening, or substitution. During treatment, three types of visual biofeedback were used as intervention methods, achieving favorable therapeutic outcomes. All patients were treated for 4 to 6 weeks, resulting in complete recovery in 63 cases and improvement in 5 cases; 0 cases showed no improvement. Among postoperative palatoplasty patients with good velopharyngeal function, alveolar affricates (/ts/) show the highest error rates in both Han and ethnic minority patients. Providing targeted foundational training before inducing the target sound and utilizing three visual biofeedback interventions in rotation yields optimal results. The third biofeedback approach showed particularly high acceptance among patients, making it especially suitable for remote therapy.

  PublisherDOI

Frequent coauthors

• Liwei Xie

  Guangdong Academy of Sciences

  61 shared

• Xudong Mai

  Zhujiang Hospital

  34 shared

• Jia Sun

  Southern Medical University

  33 shared

• Yulong Yin

  Hunan Agricultural University

  28 shared

• Bingdong Liu

  Guangdong Academy of Sciences

  26 shared

• Ye Tian

  25 shared

• Shujie Chen

  University of Sydney

  25 shared

• Xiang‐Ping Yao

  First Affiliated Hospital of Fujian Medical University

  25 shared

Awards & honors

• Alicia Cascallar Award (NCME, 2022)
• Excellent Reviewer Award (JEBS, 2020, 2023, 2024)
• UIUC List of Teachers Ranked as Excellent by Students (SP 20…