Gravar-mail: Computational optimization of associative learning experiments