Vi­su­al­iz­ing chess po­si­tion rep­re­sen­ta­tions learned using con­vo­lu­tion­al au­toen­coders