I have been a Full Professor at the Department of Computer Science, the University of Warwick since 2023, where I also serve as Head of the AI/ML Systems (AMS) Division. Prior to that I was a Senior Research Scientist at Samsung AI Centre Cambridge, where I led research on Automated Machine Learning (AutoML) for on-device intelligence.
I am a Fellow of the Alan Turing Institute, the UK’s national institute for data science and AI. At the Turing, I am also an Independent Scientific Advisor for the BridgeAI programme, and a member of the Turing Research Ethics (TREx) panel.
We have multiple openings for PhD/internship positions. Scholarships available from various sources at Warwick. Please make contact for further information.
2026/02: Two papers (on multi-condition fusion in diffusion and gloss-free sign language translation) accepted to CVPR.
2026/01: Three papers (on diffusion-based trajectory recovery, adaptive traffic signal control, and unified data platform with LLMs) accepted to WWW.
2025/09: Three papers (on efficient distiallation for diffusion models, gloss-free sign languange translation, and individualized, voxel-level brain parcellations from fMRI) accepted to NeurIPS.
2025/05: Two papers(one on diffusion-based road network generation, one on dynamic graph modelling) accepted to IJCAI.
2025/05: FlexControl (computation-aware controlled T2I generation with differentiable router) accepted to ICML.
2025/04: New work on improving efficiency of diffusion models via attention reuse accepted to IJCV.
2025/01: New paper on differentiable pruning at initialisation acceepted to ICLR.
2024/12: EX-Gaze to appear in IEEE TVCG.
2024/11: AdaFlow received Best Paper Runner-up Award at SenSys 2024.
2024/10: New work on event-based eye tracking to appear in IEEE T-PAMI.
2024/10: New paper on on-face gesture recognition via on-ear vibration sensing accepted to IEEE TMC.
2024/09: New paper on opportunistic inference accepted to SenSys 2024.
2024/07: New paper on trajectory recovery accepted to CIKM 2024.
2024/05: New paper on hierarchical generative NAS accepted to ICML 2024.
2024/02: New paper on graph-based sign language processing accepted to CVPR 2024.
2023/10: New work on rotational speed estimation with event camera to appear in IEEE TMC.
2023/09: New EV-Eye dataset available on GitHub.
2023/09: Two papers accepted to NeurIPS 2023.
2023/09: T-CET received Best Paper Award at AutoML Conf 2023.
2023/07: New work on generating neural architectures accepted to AutoML Conf.
2023/05: New work on training-free NAS (T-CET) accepted to AutoML Conf.
2022/12: New paper on fleet management for shared e-mobility systems to appear in IEEE T-ITS.
2022/11: New paper on efficient NAS accepted to AAAI 2023.
2022/09: Our paper on macro NAS benchmarks (BLOX) accepted to NeurIPS 2022.
2022/08: Our teams won both the 1st and 2nd places at the Zero Cost NAS Competition held in conjunction with the AutoML’22 conference.
2022/06: Received the National AI Strategy Fund to support research on Macro Neural Architecture Search.
2022/03: New work on spatial data synthesis: generating synthetic spatial data with local differential privacy.
2021/09: New paper on deployment optimisation for electric vehicle charging infrastructure to appear in IEEE T-ITS.
2021/08: Our work on blind video super-resolution accepted to ICCVW.
2021/07: New work on zero-cost differentiable NAS: 25min search on DARTS CNN space with a single GPU.
2021/06: Paper on vibration-based communication between IoT devices to appear in ACM TOSN.
2021/01: New work on representation learning for event streams to appear in IEEE TPAMI.
2020/08: Our work on NAS for tiny perceptual super-resolution accepted to ECCV.
2020/06: New work on semantic-instance segmentation for point clouds accepted to ICRA.
2020/05: New work on GCN-based spatio-temporal demand prediction for electric vehicle sharing in urban settings to appear in ACM IMWUT.
2020/04: Our work on secured cyber-physical social interactions on wearable devices to appear in ACM TOSN.
2020/02: Our work on fleet rebalancing using deep reinforcement learning accepted to IJCAI.