Xiaowen Dong
Associate Professor of Engineering Science
Develops signal processing and machine learning for graph-structured data, with applications to social, urban and financial networks.
Xiaowen Dong's lab at Oxford works on signal processing & information theory and speech enhancement & signal processing. 71 papers since 2023, 29 at top venues, mostly ICLR and NeurIPS. 15 PhD students and 6 master's students. 12 alumni. Applications close .
At a glance
- papers, 2023–2026
- 19/ year
- at T1 venues
- 41%
- PhD students
- 15
- MS students
- 6
- alumni listed
- 12
Papers by yearfrom the lab homepage, 2023–2026
- 14 papers2026 · T1 7 · collecting
- 19 papers2025 · T1 10
- 20 papers2024 · T1 5
- 18 papers2023 · T1 7
T1 venuesOtherVenues: ICLR, NeurIPS, ICASSP, AISTATS, ICML
Research focuspaper titles that use each topic's words
- financial3 papers
- urban1 paper
- Signal processing & information theory0 papers
- Speech enhancement & signal processing0 papers
- develops signal0 papers
Paper list
From the lab homepage.
202614 papers
- TMLR (T1)A survey on over-smoothing and over-squashing: Unified propagation perspectives on graph neural networks and transformers (opens in a new tab)
- ICLR (T1)Attention sinks and compression valleys in LLMs are two sides of the same coin (opens in a new tab)
- ICLR (T1)Bures-Wasserstein flow matching for graph generation (opens in a new tab)
- ICLR WorkshopCan graph foundation models generalize over architecture? (opens in a new tab)
- ICASSP (T1)Data-driven graph filters via adaptive spectral shaping (opens in a new tab)
- ICLR (T1)gLSTM: Mitigating over-squashing by increasing storage capacity (opens in a new tab)
- EGUInferring daily heat-related mortality risk from weekly death records with machine learning
- ICLR Workshopk-Maximum inner product attention for graph transformers and the expressive power of GraphGPS (opens in a new tab)
- NetSciNetwork topology determines when small AI models collectively outperform a single larger model
- IEEE Transactions on Signal ProcessingOn the impact of sample size in reconstructing noisy graph signals: A theoretical characterisation (opens in a new tab)
- ICLR WorkshopScalable message passing neural networks: No need for attention in large graph representation learning (opens in a new tab)
- AISTATS (T1)Spectral clustering for directed graphs via likelihood estimation on stochastic block models (opens in a new tab)
202519 papers
- ICLR (T1)Bundle neural networks for message diffusion on graphs (opens in a new tab)
- LoGFLORA: GNNs as predictors of agentic workflow performances (opens in a new tab)
- International Journal of ForecastingForecasting realized volatility with spillover effects: Perspectives from graph neural networks (opens in a new tab)
- AISTATS (T1)Heterogeneous graph structure learning through the lens of data-generating processes (opens in a new tab)
- AAAI WorkshopJudge a book by its cover: Investigating multi-modal LLMs for multi-page handwritten document transcription
- NeurIPS WorkshopLGDC: Latent graph diffusion via spectrum-preserving coarsening
- ProbNumNatural evolutionary search meets probabilistic numerics (opens in a new tab)
- ICLR (T1)Neural spacetimes for DAG representation learning (opens in a new tab)
- ICML (T1)On measuring long-range interactions in graph neural networks (opens in a new tab)
- CAMSAPOn the impact of downstream tasks on sampling and reconstructing noisy graph signals (opens in a new tab)
- NeurIPS WorkshopOn the importance of task complexity in evaluating LLM-based multi-agent systems
- NeurIPS (T1)On the stability of graph convolutional neural networks: A probabilistic perspective (opens in a new tab)
202420 papers
- ICLR (T1)A characterization theorem for equivariant networks with point-wise activations (opens in a new tab)
- GSPA novel smoothness prior for hypergraph machine learning
- NeurIPS (T1)Bayesian optimization of functions over node subsets in graphs (opens in a new tab)
- GSPDecentralized and lifelong-adaptive multi-agent collaborative learning
- ICAIFExtracting alpha from financial analyst networks (opens in a new tab)
- NeurIPS WorkshopGraph classification Gaussian processes via Hodgelet spectral features
- Machine LearningGraph similarity learning for change-point detection in dynamic networks (opens in a new tab)
- Journal of Financial Econometrics, article no. nbae026Graph-based methods for forecasting realized covariances (opens in a new tab)
- GSPHeterogeneous graph structure learning: A statistical perspective
- ICASSP (T1)Hypergraph transformer for semi-supervised classification (opens in a new tab)
- ICASSP (T1)Hypergraph-MLP: Learning on hypergraphs without message passing (opens in a new tab)
- Journal of Portfolio ManagementLarge language models for financial and investment management: Applications and benchmarks (opens in a new tab)
202318 papers
- NeurIPS (T1)Bayesian optimisation of functions on graphs (opens in a new tab)
- Nature Communications (T1)Behavioral changes during the COVID-19 pandemic decreased income diversity of urban encounters (opens in a new tab)
- ICML (T1)DRew: Dynamically rewired message passing with delay (opens in a new tab)
- IEEE Transactions on Signal and Information Processing over NetworksGaussian processes on graphs via spectral kernel learning (opens in a new tab)
- UAI (T1)Graph classification Gaussian processes via spectral features (opens in a new tab)
- ICML WorkshopGromov-Hausdorff distances for comparing product manifolds of model spaces
- GSPLearning heterogeneous graphs with generalized smoothness
- ICASSP (T1)Learning hypergraphs from signals with dual smoothness prior (opens in a new tab)
- ICAIF Workshop on NLP and Network Analysis in Financial ApplicationsLearning to learn network momentum
- Machine LearningLocal2Global: A distributed approach for scaling representation learning on graphs (opens in a new tab)
- International Conference on Complex Networks and Their ApplicationsMaximum likelihood estimation on stochastic blockmodels for directed graph clustering
- NeurIPS (T1)Neural latent geometry search: Product manifold inference via Gromov-Hausdorff-informed Bayesian optimization (opens in a new tab)
Where alumni went12 listed, 0% to industry
Industry 0Academia 3
Also to
- NTU Singapore
- Oxford Big Data Institute
- University of Lugano
Apply to Oxford2027-28
DPhil in Computer Science
- Round
- 2027-28 entry (January round)
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNTsinghua
Yuxiao Dong
- Graph mining & networks
- GBCambridge
Hatice Gunes
- Social computing & responsible AI
- USUSC
Salman Avestimehr
- Trustworthy & safe ML
- USUT Austin
Joydeep Ghosh
- Trustworthy & safe ML
- HKCUHK
Irwin King
- Graph learning
- GBUCL
Miguel Rodrigues
- Learning theory & optimization
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Oxfordcloses