In this section
Most of the members of this group are from the Statistics Section and Biomaths research group of the Department of Mathematics. Below you can find a list of research areas that members of this group are currently working on and/or would like to work on by applying their developed mathematical and statistical methods.
@article{Ng:2026:10.1126/sciadv.aec6546,
author = {Ng, WK and Dranczewski, J and Fischer, A and Raziman, TV and Saxena, D and Farchy, T and Stenning, K and Peters, J and Schmid, H and Branford, WR and Barahona, M and Moselund, K and Sapienza, R and Gartside, JC},
doi = {10.1126/sciadv.aec6546},
journal = {Science Advances},
title = {Few-shot neuromorphic vision in a nonlinear photonic network laser},
url = {http://dx.doi.org/10.1126/sciadv.aec6546},
volume = {12},
year = {2026}
}
TY - JOUR
AB - With the growing prevalence of AI, demand increases for hardware that mimics the brain’s ability to extract structure from limited data. In the retina, ganglion cells detect features from sparse inputs via lateral inhibition, where neurons antagonistically suppress activity of neighbouring cells. Biological neurons exhibit diverse heterogeneous nonlinear responses, linked to robustlearning and strong performance in low-data regimes.Here, we introduce a retinally-inspired photonic computing system where spatially-competing lasing modes in a random network laser act as heterogeneous, inhibitively-coupled neurons- enabling feature detection, few-shot classification, and segmentation. This silicon-compatible scheme harnesses heterogeneous excitatory and inhibitory nonlinear physical dynamics which give rise to emergent photonic computing behaviour, including parallel feature detection and strong performance when training data is scarce. We report 98.05% and 87.85% accuracy on MNIST and Fashion-MNIST, and 90.12% on BreaKHis cancer diagnosis- outperforming software CNNs including EfficientNetV2 and the vision transformer ViT in few-shot and class-imbalanced regimes with training sets of up to several hundred images. We demonstrate combined segmentation and classification on the HAM10k skin lesion dataset, achieving DICE and Jaccard scores of 84.49% and 74.80%. These results demonstratethe potential of random lasing networks as nonlinear photonic learning systems, and highlight the ability of heterogeneous nonlinear dynamics to support strong learning in challenging low-data scenarios.
AU - Ng,WK
AU - Dranczewski,J
AU - Fischer,A
AU - Raziman,TV
AU - Saxena,D
AU - Farchy,T
AU - Stenning,K
AU - Peters,J
AU - Schmid,H
AU - Branford,WR
AU - Barahona,M
AU - Moselund,K
AU - Sapienza,R
AU - Gartside,JC
DO - 10.1126/sciadv.aec6546
PY - 2026///
SN - 2375-2548
TI - Few-shot neuromorphic vision in a nonlinear photonic network laser
T2 - Science Advances
UR - http://dx.doi.org/10.1126/sciadv.aec6546
UR - https://www.science.org/doi/full/10.1126/sciadv.aec6546
VL - 12
ER -
If you are interested in meeting with members of the group please contact Marina Evangelou