B. Silvano Zanutto

Constructing neural networks with pre-specified dynamics

This paper introduces a method to construct binary recurrent neural networks with pre-specified dynamics and demonstrates its application in modeling various types of network behavior.

Probing the structure-function relationship with neural networks constructed by solving a system of linear equations

The paper introduces a method to construct neural networks by solving a system of linear equations to probe structure-function relationships in a memory task, finding networks that traditional optimization algorithms fail to produce.

Exploring the limits of learning: segregation of information integration and response selection is required for learning a serial reversal task

This paper shows that a single neural population cannot learn a serial reversal task due to limitations in integrating information and selecting responses, proposing a model that separates these functions.