Image preprocessor based on thalamo-cortical neurons
The project concerns the construction of a model of the thalamus, which is the part of the diencephalon. Various models of thalamo-cortical neurons are used in the studies. The rich functionality of the thalamus is the inspiration to create a preprocessor for sensor and vision data.
The project additionally adapts the cusp catastrophe model used in psychiatry to model multidimensional behavior. The model learning task is based on observations of infant learning processes.
The costs of hardware implementation of the AI model using FPGA matrices are also considered. An example area of application of the model is the implementation of self-learning agents with severely limited resource availability.
The project results were published, among others, in the NATO Science for Peace and Security Series.
The project is implemented in cooperation with Universidade NOVA de Lisboa.
Funding
Statutory Activities for Young Staff No. 0311/SBAD/0714
Publications
Explainable spiking neural network for real time feature classification / Szymon Szczęsny, Damian Huderek, Łukasz Przyborowski // Journal of Experimental & Theoretical Artificial Intelligence – 2022, vol. 35, no. 1, s. 77-92
Spiking Neural Network with Linear Computational Complexity for Waveform Analysis in Amperometry / Szymon Szczęsny, Damian Huderek, Łukasz Przyborowski // Sensors – 2021, vol. 21, no. 9, s. 3276-1-3276-16
3rd Generation Networks in UAV Vision Systems / Szymon Szczęsny, Damian Huderek // W: Modern Technologies Enabling Safe and Secure UAV Operation in Urban Airspace / red. Paweł Śniatała, Sundararaja Sitharama Iyengar, Amine Bendarma, Maciej Klósak – Amsterdam, Netherlands : IOS Press, 2021 – s. 107-114
