Regina Zhang

Regina Zhang

  • Postdoctoral Associate, Yale UniversityRadiology & Biomedical Imaging, with Xiaofeng Liu
  • Postdoctoral Researcher, University of Cambridgewith Pietro Liò

I develop AI methods for data science and biomedical informatics, with a focus on learning from graphs and time series that are noisy, incomplete, and evolving. My current work designs algorithms to forecast how Alzheimer’s pathology progresses across the brain and to advance PET tracer research.

Google Scholar (opens in a new tab) reginazhang955@gmail.com
Regina Zhang

News

  • SPIE Medical Imaging 2027 — first-authored paper on tau PET forecasting with adaptive normative connectivity priors accepted as an oral presentation (Vancouver). paper
  • The VLDB Journal — first-authored paper Efficient Prompt Learning for Traffic Forecasting accepted. paper
  • JCTC — our survey and benchmark of deep learning for molecular property prediction published as a cover article (Vol. 22, Issue 10). paper
  • ACM TOIS — first-authored paper HMamba: Hyperbolic Mamba for Sequential Recommendation accepted (published 23 May 2026). paper
  • IEEE TKDE — first-authored paper M²Rec: Multi-scale Mamba for Sequential Recommendation accepted. paper

Research frameworkGraph and temporal data connect through shared AI methods to data science and biomedical informatics. Surveys and benchmarks support the research foundation. Venue labels link to the papers.A shared graph and foundation model foundation across data science and biomedical informaticsLearning from noisy, incomplete and evolving graphs and time series01 DATA02 AI METHODS03 IMPACTstSpatiotemporal graphsSensors, urban networks,trajectories and time seriesMolecular graphsAtoms, bonds andmolecular propertiesBrain networks & PETRegional connectivity andlongitudinal tau imagingRobust graph learningContrastive learning and learned augmentationfor noisy and incomplete graph dataICML’23WWW’23CIKM’25ICDE’24EMNLP’25Efficient temporal modelingDistillation, attention, prompts and state-spacemodels for scalable sequence learningAAAI’25 OralICDE’26VLDBJ’26TKDE’26TOIS’26ICDE’23TKDD’24SDM’22Knowledge-guided biomedical AIConnectivity and chemical priors, foundationmodels and calibrated forecastsJCTC’26 CoverCVPR-W’26CVPR-W’26SPIE MI’27 OralTau PET forecastingData scienceTraffic and urban forecastingTrajectory and sports analyticsPOI and sequential recommendationBiomedical informaticst₀t₁t₂Brain regions = graph nodesAlzheimer’s tau progressionPET tracer researchTumor segmentation · molecular properties04 SURVEYS & BENCHMARKSMolecular propertyprediction benchmarksJCTC’26 CoverPoint-of-interestrecommendationTKDE’25Generative models forspatiotemporal dataDSE’26Time-series analysisin the frequency domainPreprintLLMLLM-driven time seriesfrom prompts to agentsPreprintEfficiency of LLM codeEffiBench-XNeurIPS’25LLM agents forsoftware engineeringPreprint
Fig. 1 | Research framework. 01, Graph and temporal data from data science and biomedical informatics. 02, Shared methods address robustness, efficiency and domain knowledge. 03, Applications connect the method foundation to data science and biomedical informatics. 04, Surveys and benchmarks consolidate these areas. Venue labels link to the papers below.

PublicationsGoogle Scholar

Each paper has a simplified visual overview of its method or dataset, collapsed by default. Use “Show overview” on a paper, or “Show all overviews”, and click a figure to enlarge it.

Venue rankings: CCF China Computer Federation recommended list (7th edition, 2026); CORE ICORE conference ranking (2026); IF 2025 two-year Journal Impact Factor (2026 release).

Journal articles

  1. SURVEY FOCUSMolecular properties

    A Systematic Survey and Benchmarks of Deep Learning for Molecular Property Prediction in the Foundation Model Era (opens in a new tab)

    Zongru Li et al.

    Journal of Chemical Theory and Computation, 2026. Cover article

    Corresponding authorIF 5.8

  2. RESEARCH TARGETTraffic forecasting

    Efficient Prompt Learning for Traffic Prediction

    Regina Zhang et al.

    The VLDB Journal, 2026

    First authorCCF AIF 5.3

  3. RESEARCH TARGETNext-item recommendation

    M²Rec: Multi-scale Mamba for Sequential Recommendation

    Regina Zhang et al.

    IEEE Transactions on Knowledge and Data Engineering, 2026

    First authorCCF AIF 11.6

  4. RESEARCH TARGETNext-item recommendation

    HMamba: Hyperbolic Mamba for Sequential Recommendation

    Regina Zhang et al.

    ACM Transactions on Information Systems, 2026

    First authorCCF AIF 11.2

  5. SURVEY FOCUSSpatiotemporal data mining

    A Survey of Generative Techniques for Spatial-Temporal Data Mining

    Regina Zhang et al.

    Data Science and Engineering, 2026

    First authorIF 4.1

  6. SURVEY FOCUSNext-place recommendation

    A Survey on Point-of-Interest Recommendation: Models, Architectures, and Security (opens in a new tab)

    Regina Zhang et al.

    IEEE Transactions on Knowledge and Data Engineering, 2025

    First authorCCF AIF 11.6

  7. DATASET TARGETBilliards layouts & trajectories

    Billiards Sports Analytics: Datasets and Tasks (opens in a new tab)

    Regina Zhang et al.

    ACM Transactions on Knowledge Discovery from Data, 2024

    First authorCCF BIF 6.9

Conference papers

  1. tau PET t₀forecast t₁
    RESEARCH TARGETLongitudinal tau PET forecasting

    Adaptive Normative Connectivity Priors for Longitudinal Tau PET Forecasting with Multi-Relation Graph Neural Networks

    Regina Zhang, Jonghye Woo, Georges El Fakhri, Jinsong Ouyang, Xiaofeng Liu

    SPIE Medical Imaging, Computer-Aided Diagnosis, 2027

    First authorOral

  2. RESEARCH TARGETTime-series forecasting

    AutoHFormer: Efficient Hierarchical Autoregressive Transformer for Time Series Prediction (opens in a new tab)

    Regina Zhang et al.

    IEEE International Conference on Data Engineering (ICDE), 2026

    First authorCCF ACORE A*

  3. RESEARCH TARGETTraffic forecasting

    Efficient Traffic Prediction Through Spatio-Temporal Distillation (opens in a new tab)

    Regina Zhang et al.

    AAAI Conference on Artificial Intelligence, 2025. Oral

    First authorCCF ACORE A*

  4. RESEARCH TARGETUrban region representations

    HGAurban: Heterogeneous Graph Autoencoding for Urban Spatial-Temporal Learning (opens in a new tab)

    Regina Zhang et al.

    ACM International Conference on Information and Knowledge Management (CIKM), 2025

    First authorCCF BCORE A

  5. BENCHMARK TARGETGenerated-code efficiency

    EffiBench-X: A Multi-Language Benchmark for Measuring Efficiency of LLM-Generated Code (opens in a new tab)

    Regina Zhang and coauthors

    Conference on Neural Information Processing Systems (NeurIPS), 2025

    Corresponding authorCCF ACORE A*

  6. ?
    RESEARCH TARGETNatural language to SQL

    Graph-Reward-SQL: Execution-Free Reinforcement Learning for Text-to-SQL via Graph Matching and Stepwise Reward (opens in a new tab)

    With coauthors including Regina Zhang

    Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025

    CCF BCORE A*

  7. RESEARCH TARGETUser–item recommendation

    Graph Augmentation for Recommendation (opens in a new tab)

    Regina Zhang et al.

    IEEE International Conference on Data Engineering (ICDE), 2024

    First authorCCF ACORE A*

  8. RESEARCH TARGETUrban region representations

    Spatial-Temporal Graph Learning with Adversarial Contrastive Adaptation (opens in a new tab)

    Regina Zhang et al.

    International Conference on Machine Learning (ICML), 2023

    First authorCCF ACORE A*

  9. RESEARCH TARGETUrban region representations

    Automated Spatio-Temporal Graph Contrastive Learning (opens in a new tab)

    Regina Zhang et al.

    The Web Conference (WWW), 2023

    First authorCCF ACORE A*

  10. !
    RESEARCH TARGETAnomalous trip segments

    Online Anomalous Subtrajectory Detection with Road Network Enhanced Reinforcement Learning (opens in a new tab)

    Regina Zhang et al.

    IEEE International Conference on Data Engineering (ICDE), 2023

    First authorCCF ACORE A*

  11. RESEARCH TARGETBreak-shot quality & design

    On Predicting and Generating a Good Break Shot in Billiards Sports (opens in a new tab)

    Regina Zhang et al.

    SIAM International Conference on Data Mining (SDM), 2022

    First authorCCF BCORE A

Workshop papers

  1. PET / CTLESIONS
    RESEARCH TARGETWhole-body tumor segmentation

    An Open Multi-Center Whole-Body FDG PET/CT Foundation Model for Tumor Segmentation (opens in a new tab)

    Xiaofeng Liu, Regina Zhang, et al.

    CVPR Workshop on Multimodal Foundation Models for Biomedicine, 2026

  2. t₀t₁?
    RESEARCH TARGETChange-aware medical VQA

    Longitudinal Medical Visual Question Answering Guided by Vision Foundation Models for Consistent Attention

    J. Wu, Regina Zhang, et al.

    CVPR PHAROS Workshop on AI for Medical Imaging and Healthcare, 2026

Preprints

  1. SURVEY FOCUSFrequency-domain time series

    Time Series Analysis in Frequency Domain: A Survey of Open Challenges, Opportunities and Benchmarks (opens in a new tab)

    Regina Zhang et al.

    Preprint, under revision

    First author

  2. LLM
    SURVEY FOCUSLLM-based time-series analysis

    From Prompts to Agents: A Comprehensive Survey of LLM-Driven Time Series Analysis (opens in a new tab)

    Regina Zhang et al.

    Preprint, under revision

    First author

  3. SURVEY FOCUSSoftware engineering agents

    A Comprehensive Survey on Benchmarks and Solutions in Software Engineering of LLM-Empowered Agentic System (opens in a new tab)

    Regina Zhang and coauthors

    Preprint, under review

    Corresponding author

Teaching, talks & service

Teaching

Teaching Assistant, FITE7405: Techniques in Computational Finance, The University of Hong Kong. Four offerings, 2021–2024.

Prepared to teach undergraduate machine learning and data science, and graduate courses in graph machine learning, time-series analysis, and AI for science and medicine.

Professional service

  • Symposium co-organizer, IEEE Symposium on CI in Natural Language Processing and Social Media (IEEE CI-NLPSoMe 2027, part of IEEE SSCI 2027), Gold Coast, AustraliaWith Hongzhi Yin and Tong Chen (University of Queensland), Usman Naseem (Macquarie University), and Henry Nguyen (Griffith University)
  • Workshop organizer, “Neural Reasoning”, LondonWith Pietro Liò (University of Cambridge)
  • Session Chair, Time Series, SIAM International Conference on Data Mining (SDM)

Reviewer for ICML, NeurIPS, ICLR, KDD, AAAI, The Web Conference, and IEEE TKDE.

Honors

  • Baidu Research FellowshipOne of 40 recipients worldwide
  • Student Travel Award, The Web Conference
  • Postgraduate Scholarship, The University of Hong Kong

Invited talks

  • Keynote, AAAI Workshop on Neural-Symbolic NLP and Knowledge Graph Reasoning
  • University of Cambridge
  • Yale University
  • Cornell University
  • Westlake University

Research agendaDirections for an independent group

My future research direction will focus on developing novel, robust, and efficient foundation models and graph neural networks for evolving graphs and time series, with rigorous validation in high-stakes applications, beginning with neurodegenerative disease.

  1. AIM 1 · FORECAST
    PET + brain graphCalibrated forecast

    Forecasting disease progression on brain networks

    Forecast regional tau accumulation in Alzheimer’s disease from longitudinal PET, using brain connectivity as structure, accounting for scan quality, and producing calibrated uncertainty validated across cohorts.

  2. AIM 2 · PERSONALIZE
    Imaging + genomicsIndividual forecasts

    Individualized models of progression

    Integrate imaging with genomic information so that forecasts reflect the individual patient rather than the population average.

  3. AIM 3 · GENERALIZE
    RobustnessTransferEfficiencyFoundation modelsOpen benchmarks

    Reusable methods and rigorous benchmarks

    Release algorithms and benchmarks that show when robust and efficient graph learning helps, how it transfers across domains, and what it costs.

Contact

reginazhang955@gmail.com

Department of Radiology & Biomedical Imaging, Yale School of Medicine, New Haven, Connecticut.