Recently, Orhan and Lake demonstrated the computational plausibility that children can acquire sophisticated visual representations from natural input data without inherent biases, challenging the need for innate constraints in human learning. The findings may also reveal crucial properties of early visual learning and inform theories of human visual development.

中文翻译：

---

超越可学习性：利用 DNN 了解人类视觉发育


最近，奥尔罕和莱克证明了儿童可以从自然输入数据中获取复杂的视觉表示而没有固有偏见的计算合理性，挑战了人类学习中固有限制的需要。这些发现还可能揭示早期视觉学习的关键特性，并为人类视觉发育的理论提供信息。