Samuel Chun Hei Lam
University of Oxford
Andrew Wiles Building
Radcliffe Observatory Quarter
Woodstock Road
Oxford
OX2 6GG
Journal Publications
S.C.H. Lam, J. Sirignano and K. Spiliopoulos. Weak Convergence Analysis of Real-time Recurrent Learning (RTRL) for a class of Recurrent Neural Networks. (2025+) Submitted. arxiv:2501.08040
S.C.H. Lam, J. Sirignano and Z. Wang. Weak Convergence Analysis of Online Neural Actor-Critic Algorithms (2024+) Resubmitted for revision.
arxiv:2403.16825
S.C.H. Lam, J. Sirignano and K. Spiliopoulos. Kernel Limit of Recurrent Neural Networks Trained on Ergodic Data Sequence (2024+) Resubmitted for revision.
arxiv:2308.14555
Conference Publication
I. Price, N.D. Ball, S.C.H. Lam, A.C. Jones, J. Tanner. Deep Neural Network Initialization with Sparsity Inducing Activations. (2024, ICLR)
arXiv:2402.16184
Samuel's research interests are broadly theories of deep learning. He is currently researching on the kernel-limit and mean-field analyses of neural networks.
Samuel is a DPhil student under the EPSRC CDT Mathematics of Random System, jointly offered by the University of Oxford and Imperial College London. Prior to his DPhil studies, he read Mathematics as an undergraduate at Imperial College London, spending a year abroad at the Massachusetts Institute of Technology (MIT).
- MCF Deep Learning (Tutor, Hilary 2024, 2025; TA, Hilary 2023)
CDT studentship, jointly funded by the EPRSC and the Oxford-Radcliffe Graduate Scholarship. (2022-2026)
Institute of Mathematics and its Applications (IMA) Prize, Imperial College London. (2022)
Candidate for MIT-Imperial Exchange Programme (2020-2021)