视觉表征的通用维度
Universal dimensions of visual representation
AI总结：
该研究通过分析数十万视觉网络表征维度，发现不同架构和任务的网络会收敛学习到一组通用图像表征，且这些通用维度最能解释其与人脑表征的一致性，揭示了人工与生物视觉相似性的核心来源。
摘要：
视觉神经网络模型之所以能学习到与大脑一致的表征，是因为它们与生物视觉共享架构约束和任务目标，还是因为它们学习到了自然图像处理的通用特征？我们对来自构造各异的视觉神经网络的数十万个表征维度进行了通用性刻画。我们发现，尽管架构和任务目标不同的网络在表面层面看起来高度不同，但它们会使用一组共享的潜在维度来表示自然图像。接着，通过将这些网络与用fMRI测量到的人脑表征进行比较，我们发现神经网络中最能与大脑对齐的表征，是那些具有通用性且不依赖于网络具体特征的表征。值得注意的是，每个网络都可以被缩减到少于十个最通用的维度，而对其与人脑的表征相似性几乎没有影响。这些结果表明，人工视觉与生物视觉之间的深层相似性，主要由一组核心的通用图像表征所支配，这些表征被不同系统收敛地学习到。
英文摘要：
Do neural network models of vision learn brain-aligned representations because they share architectural constraints and task objectives with biological vision or because they learn universal features of natural image processing? We characterized the universality of hundreds of thousands of representational dimensions from visual neural networks with varied construction. We found that networks with varied architectures and task objectives learn to represent natural images using a shared set of latent dimensions, despite appearing highly distinct at a surface level. Next, by comparing these networks with human brain representations measured with fMRI, we found that the most brain-aligned representations in neural networks are those that are universal and independent of a network's specific characteristics. Remarkably, each network can be reduced to fewer than ten of its most universal dimensions with little impact on its representational similarity to the human brain. These results suggest that the underlying similarities between artificial and biological vision are primarily governed by a core set of universal image representations that are convergently learned by diverse systems.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
