16 January. 2021
Switch neurons (Vassiliades & Christodoulou, 2016) were proposed as a mechanism that enables flexible and adaptive behavior in an artificial neural network (ANN) through gating the flow of information. These models have shown to produce optimal exploration strategies in certain tasks, however, their full potential currently remains unexplored. We will be looking into projects related to 1) learning such ANN architectures through neuroevolution / neural architecture search; 2) extending such neurons to i) handle noisy signals, ii) make them more adaptive; 3) using them in more complex environments; 4) training them using gradient-descent; 5) comparing them with other ANNs.
References:
Vassiliades, V. and Christodoulou, C. (2016). Behavioral Plasticity Through the Modulation of Switch Neurons. Neural Networks, 74: 35-51.
Good programming skills (C++ or Python, multi-processing, ability to adapt existing code). Knowledge of machine learning, neural and evolutionary computation will be considered an advantage.