Zipeng Wu
Temporal representation learning for foundation / world models and non-stationary time-series ML.
University of Birmingham
About me
I am a PhD researcher in Applied Mathematics at the University of Birmingham. My current research studies how temporal structure can be represented, decomposed, compared, and evaluated across time series, action histories, VLA trajectories, model rollouts, temporal memory, and noisy sequential systems.
I am especially interested in temporal/time-series representation as a bridge between classical time-series machine learning and emerging foundation or world-model systems. Recent work includes decomposition as representation extraction, stationarity-aware retrieval, time-series diagnostics, and language-action temporal tokenization.
Current Research Focuses
(1) Temporal representation for foundation and world models, including model rollouts, action histories, temporal memory, and VLA trajectories.
(2) Time-series representation learning under non-stationarity, including decomposition, retrieval, forecasting diagnostics, and robust evaluation.
(3) Representation-driven research software for inspectable temporal components, compact JSON outputs, and shareable reports.
news
| Jul 2026 | UKRI/AIRR Gateway Project Language-Action Time-Series Tokenization for Efficient VLA Policies allocated 10,000 GPUHR on Isambard-AI, with nominal compute-resource value GBP 45,000; compute resources only, not direct cash funding. |
|---|---|
| Jun 2026 | UKRI/AIRR Gateway Project Time Series Language and Foundation Model allocated 10,000 GPUHR on Isambard-AI, with nominal compute-resource value GBP 45,000; compute resources only, not direct cash funding. |
| Jun 2026 | Co-authored paper accepted to KDD 2026, CORE/ICORE A*. |
| May 2026 | First-author paper accepted to ICML 2026, CORE/ICORE A*. |