Modeling Retinal Ganglion Cells with Neural Differential Equations (Student Abstract)

Kacper Dobek; Daniel Jankowski; Krzysztof Krawiec

AAAI 2026technical0 citations

Modeling Retinal Ganglion Cells with Neural Differential Equations (Student Abstract)

Kacper Dobek, Daniel Jankowski, Krzysztof Krawiec

Abstract

This work explores Liquid Time-Constant Networks (LTCs) and Closed-form Continuous-time Networks (CfCs) for modeling retinal ganglion cell activity in tiger salamanders across three datasets. Compared to a convolutional baseline and an LSTM, both architectures achieved lower MAE, faster convergence, smaller model sizes, and favorable query times, though with slightly lower Pearson correlation. Their efficiency and adaptability make them well suited for scenarios with limited data and frequent retraining, such as edge deployments in vision prosthetics.

BibTeX

@inproceedings{aaai2026_modelingretinalg,
  title = {Modeling Retinal Ganglion Cells with Neural Differential Equations (Student Abstract)},
  author = {Kacper Dobek and Daniel Jankowski and Krzysztof Krawiec},
  booktitle = {AAAI 2026},
  year = {2026}
}