Learning in neural spatial interaction models: A statistical perspective

 In this paper we view learning as an unconstrained non-linear minimization problem in which the objective function is defined by the negative log-li