New Book on AI Futures
Co-authored AI全景探索: 人工智能的未来之旅, offering a multidisciplinary perspective on responsible AI and energy innovation (Mandarin edition, 2024).
About
I am a Senior Researcher at the Technical University of Denmark, focusing on trustworthy data platforms and AI applications for accelerating the energy transition. My research bridges smart meter analytics, intelligent energy systems, and evidence-based policy for climate-neutral cities.
I received my Ph.D. in Computer Science from Aalborg University and have held research positions at Åbo Akademi University, the University of Waterloo, and IBM Toronto Research. My work combines large-scale data platforms, machine learning, and behavioral insights to support utilities and policy makers with actionable intelligence.
My research interests span smart meter data analytics, scalable data management, IoT platforms, and privacy-preserving machine learning for energy systems. I actively collaborate with public authorities, industry partners, and interdisciplinary research teams across Europe, North America, and Asia.
Research Areas
Smart Meter Analytics
Developing scalable data platforms and benchmarking systems for smart meter data analysis, enabling utilities to extract actionable insights from large-scale energy consumption data.
Machine Learning for Energy
Advancing privacy-preserving federated learning, anomaly detection, and AI robustness for energy theft detection, demand forecasting, and electrification insights.
IoT & Data Platforms
Designing trustworthy IoT platforms that connect behavioral and technical energy data, supporting decision-making for municipalities and distribution system operators.
Urban Sustainability
Creating evidence-based tools for climate-neutral cities, including behavioral models for air quality, carbon reduction strategies, and industrial energy flexibility markets.
Recent Highlights
Privacy-Preserving Energy Analytics
Published heterogeneous federated learning framework for electricity theft detection in Applied Energy, enabling utilities to collaborate without sharing raw data.
EU H2020 ClairCity Project
Delivered behavioral models and interactive tools that informed urban carbon and air quality strategies across six European cities.
Selected Projects
A research-driven, modular PyTorch framework for advanced time series analysis, excelling in multi-source and sparse data scenarios. Features LLM-inspired architectures, Variational Autoencoders, and classical models. ⭐ 53 stars, 7 forks on GitHub.
OPTIX: Optimising Positive-Energy Districts
ActiveOptimising Positive-Energy Districts through interoperable Digital Platforms. EU project, Period: Dec 2024 - Dec 2027.
ANSWER: Wind-Solar Energy Prediction
ActiveA heterogeneous distributed prediction model for wind-solar energy production. Period: Apr 2024 - Mar 2026.
S3EM: Semantic 3D Energy Model
MSCAMarie Skłodowska-Curie Actions project developing a Semantic 3D Energy Model for building-integrated photovoltaics (BIPV) systems. Uses multimodal sensing and reinforcement learning to map solar potential on rooftops and facades, optimise BIPV placement, and coordinate with other energy systems for climate-positive prosumer districts.
BEGONIA: Cross-border Operational Digital Platforms
Horizon Europe CSAEU Coordination and Support Action accelerating digital transformation across energy and transport via interoperable Operational Digital Platforms. BEGONIA mobilises European data, cloud, edge, and connectivity assets to standardise cross-border processes, integrate renewables, and decarbonise mobility through DTU-led partnerships spanning Denmark, Spain, Greece, Belgium, and Austria.
RE-INTEGRATE: Transdisciplinary Climate-Compatible Energy Strategies
Climate, Energy & MobilityHorizon Europe action aligning EU–African Union energy planning ecosystems through inclusive modelling toolkits, local capacity building, and context-specific climate-compatible pathways. Deploys 3E models across eight AU contexts to co-create credible, independent strategies for integrated energy, climate, and socioeconomic development.
Selected Publications
Full publication list available on Google Scholar and ORCID. Recent highlights include:
Beyond Missing Data Imputation: Information-Theoretic Coupling of Missingness and Class Imbalance for Optimal Irregular Time Series Classification
AAAI 2026 (Oral). Presents SPECTRA combining frequency-guided encoders, missingness modeling, and prototype-based classification under irregular time series.
PolypSense3D: A Multi-Source Benchmark Dataset for Depth-Aware Polyp Size Measurement in Endoscopy
NeurIPS 2025 Datasets & Benchmarks Track. Resources: Code, Dataset DOI.
FedFree: Breaking Knowledge-sharing Barriers through Layer-wise Alignment in Heterogeneous Federated Learning
NeurIPS 2025. CC BY 4.0.
FRAME: Feature Rectification for Class Imbalance Learning
IEEE Transactions on Knowledge and Data Engineering
pyFAST: A Modular PyTorch Framework for Time Series Modeling with Multi-source and Sparse Data
arXiv preprint arXiv:2508.18891
A privacy-preserving heterogeneous federated learning framework with class imbalance learning for electricity theft detection
Applied Energy, 378, 124789
Adaptive expert fusion model for online wind power prediction
Neural Networks, 184, 107022
Temporal structure-preserving transformer for industrial load forecasting
Neural Networks, 107887
An integrated multi-criteria decision making framework for industrial excess heat recovery and utilization
Energy, 318, 134721
Temporal collaborative attention for wind power forecasting
Applied Energy, 357, 122502
Grid search with a weighted error function: Hyper-parameter optimization for financial time series forecasting
Applied Soft Computing, 154, 111362
A systematic review of data-driven approaches to fault diagnosis and early warning
Journal of Intelligent Manufacturing, 34(8), 3277-3304
Teaching & Supervision
Courses
Smart Cities (42282)
Focus on urban data infrastructures, civic innovation, and digital ethics.
Smart, Connected & Livable Cities (42280)
Studio-based course integrating IoT sensing with policy analysis.
Computer Science Foundations
Project-led pedagogy emphasizing software craftsmanship and data quality.
Academic Service
Editorial Roles
- Associate Editor, Cleaner Engineering and Technology (Elsevier)
- Guest Editor, Edge Machine Learning special issue, Applied Soft Computing
- Guest Editor, Advanced Building Energy Systems, Energies
Review Activities
Regular reviewer for IEEE TSG, IEEE Big Data, IEEE IoT Journal, Applied Energy, Energy & Buildings, Information Systems, DKE, KAIS, and Nature Scientific Reports.
Conference service for VLDB, EDBT, DaWaK, DOLAP, IoTBG, MobiSPC, EIA, IEEE BigData, and related venues.
Contact
Email
Office
Building 424, room 006
DTU Lyngby Campus
2800 Kgs. Lyngby, Denmark
Affiliation
Energy Economics & Modelling Group
Department of Technology, Management and Economics