[人人能懂AI前沿] AI的进化课：从“两种眼睛”到“内心世界”

你有没有想过，一个AI是真的学会了，还是在“假装”学好？本期节目，我们就来一场AI“认知深度”的探索之旅。我们会看到，最新的AI研究如何让机器拥有“两种眼睛”看清黑暗，学会“抄家伙”的变通，构建一张改造生物工厂的“活地图”，甚至发展出自己的“内心世界”。让我们一起揭开AI从“超级工具”走向“思考主体”的秘密。

00:00:30 告别黑暗，当摄像头拥有了两种“眼睛”

00:05:34 让机器人学会“抄家伙”的秘密

00:11:15 让AI拥有一个“内在世界”

00:17:08 AI造物，一张“活地图”如何改造生物工厂？

00:23:36 你的模型“假装”学好了吗？

本期介绍的几篇论文：

[CV] EeveeDark: A Binary Neural Framework for Low-Light Video Enhancement via Event-Guided Sensor-Level Fusion

[Hacettepe University]

https://arxiv.org/abs/2607.06217

---

[RO] FORGE: Towards Functional Tool-Use Generalization via Keypoint Trajectory Reasoning

[Nanyang Technological University & Georgia Institute of Technology]

https://arxiv.org/abs/2607.05780

---

[CL] From Application-Layer Simulation to Native Meta-Architecture: Structural Tension as an Endogenous Driver for Heterogeneous AI Evolution

[Shanghai Lixin University of Accounting and Finance]

https://arxiv.org/abs/2607.06269

---

[LG] Canopy: A Heterograph Foundation Model for Metabolic Engineering

[Twig Bio]

https://arxiv.org/abs/2607.06224

---

[CV] Association Restoration Test: Revealing Restorable Shortcuts after Unlearning

[Stanford University]

https://arxiv.org/abs/2607.05726