Sarah Marzen

The paper introduces a method using Maximum Entropy models to analyze biological sensors with bias, variance, and confusion matrices, offering computational efficiency and richer insights than mutual information.

This study shows that NMDA receptor activation is essential for forming long-term memory traces in cultured neural networks, which contribute to predictive capabilities.

This study shows humans achieve near-optimal prediction in a sequence learning task despite limited cognitive capacity, using an information-theoretic approach to measure resource-rationality.