并非所有解决方案都是平等的：深度线性神经网络中功能相似性与表征相似性的分析性分离
Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks
浏览论文内容
中文总结 AI 辅助
本研究通过分析两层线性网络和非线性模拟，发现神经网络的表征相似性与功能相似性可分离，且参数鲁棒性限制表征灵活性，表明表征对齐具有独立于功能对齐的计算意义。
中文摘要 AI 辅助
联结主义的一个基本原则是，知觉、行动和认知源于简单、相互连接的单元之间的并行计算，这些单元生成并依赖神经表征。因此，研究人员采用多变量模式分析来解码和比较人工网络与生物网络的神经代码，旨在揭示其功能。然而，关于网络表征与功能之间的关系，目前的分析性理解有限，尽管这对于任何关于潜在功能或功能相似性的量化概念都至关重要。我们使用可分析的两层线性网络和非线性网络中的数值模拟来解决这个问题。我们发现功能与表征是分离的，允许在没有功能相似性的情况下存在表征相似性，反之亦然。此外，我们表明，对输入噪声的鲁棒性和泛化误差水平都不能将表征约束到任务上。相反，对参数噪声鲁棒的网络具有有限的表征灵活性，必须采用任务特定的表征。我们的研究结果表明，表征对齐反映了超越功能对齐的计算优势，这对于解释和比较联结主义系统的表征具有重要意义。
英文摘要
A foundational principle of connectionism is that perception, action, and cognition emerge from parallel computations among simple, interconnected units that generate and rely on neural representations. Accordingly, researchers employ multivariate pattern analysis to decode and compare the neural codes of artificial and biological networks, aiming to uncover their functions. However, there is limited analytical understanding of how a network's representation and function relate, despite this being essential to any quantitative notion of underlying function or functional similarity. We address this question using analysable two-layer linear networks and numerical simulations in non-linear networks. We find that function and representation are dissociated, allowing representational similarity without functional similarity and vice versa. Further, we show that neither robustness to input noise nor the level of generalization error constrain representations to the task. In contrast, networks robust to parameter noise have limited representational flexibility and must employ task-specific representations. Our findings suggest that representational alignment reflects computational advantages beyond functional alignment alone, with significant implications for interpreting and comparing the representations of connectionist systems.
发表机构
- University of Oxford（牛津大学）
- University College London（伦敦大学学院）
机构由 AI 辅助整理，请以论文原文为准。
