[论文解读] Meaningful representations emerge from Sparse Deep Predictive Coding

Victor Boutin, Angelo Franciosini+2 more|arXiv (Cornell University)|Feb 20, 2019

Neural dynamics and brain function参考文献 17被引用 2

一句话总结

该论文提出稀疏深度预测编码（SDPC），一种分层卷积模型，结合稀疏编码与预测编码，统一解释早期视觉中反馈的神经与表征机制。结果表明，反馈连接重新组织神经交互，以支持轮廓整合并提升图像重建的抗噪能力。

ABSTRACT

Both neurophysiological and psychophysical experiments have pointed out the crucial role of recurrent and feedback connections to process context-dependent information in the early visual cortex. While numerous models have accounted for feedback effects at either neural or representational level, none of them were able to bind those two levels of analysis. Is it possible to describe feedback effects at both levels using the same model? We answer this question by combining Predictive Coding (PC) and Sparse Coding (SC) into a hierarchical and convolutional framework. In this Sparse Deep Predictive Coding (SDPC) model, the SC component models the internal recurrent processing within each layer, and the PC component describes the interactions between layers using feedforward and feedback connections. Here, we train a 2-layered SDPC on two different databases of images, and we interpret it as a model of the early visual system (V1 & V2). We first demonstrate that once the training has converged, SDPC exhibits oriented and localized receptive fields in V1 and more complex features in V2. Second, we analyze the effects of feedback on the neural organization beyond the classical receptive field of V1 neurons using interaction maps. These maps are similar to association fields and reflect the Gestalt principle of good continuation. We demonstrate that feedback signals reorganize interaction maps and modulate neural activity to promote contour integration. Third, we demonstrate at the representational level that the SDPC feedback connections are able to overcome noise in input images. Therefore, the SDPC captures the association field principle at the neural level which results in better disambiguation of blurred images at the representational level.

研究动机与目标

• 弥合早期视觉皮层中神经水平的反馈机制与表征水平的情境处理之间的差距。
• 开发一个统一模型，利用同一框架解释神经组织结构与表征鲁棒性。
• 探究反馈连接如何影响经典感受野之外的神经交互。
• 评估反馈是否在表征水平上增强对噪声或模糊视觉输入的消歧能力。

提出的方法

• 该模型在每一层内部通过稀疏编码（SC）实现递归处理，并通过预测编码（PC）实现层间前向与反馈连接。
• 在图像数据库上训练两层SDPC架构，以模拟V1与V2皮层区域。
• 通过稀疏编码建模层内递归处理，强制实现局部化与定向的感受野。
• 通过预测编码实现反馈连接，实现高层对低层活动的自上而下调制。
• 计算交互图以分析经典感受野之外的神经连接。
• 通过在模糊图像上测试重建性能，评估模型对输入噪声的鲁棒性。

实验结果

研究问题

• RQ1 单一模型能否同时解释早期视觉中反馈在神经与表征两个层面的作用？
• RQ2 反馈连接如何在经典感受野之外重新组织神经交互？
• RQ3 反馈信号在多大程度上促进轮廓整合，体现在交互图中？
• RQ4 反馈连接能否在表征水平上提升对噪声或模糊图像的消歧能力？

主要发现

• 训练后，SDPC模型在V1中发展出定向且局部化的感受野，在V2中则表现出更复杂的特征，与神经生理学观察一致。
• 反馈连接重新组织交互图，呈现出类似关联场的结构，支持良好的连续性（Gestalt）原则。
• 神经活动受反馈调制，从而增强轮廓整合，表现为交互图的空间结构。
• 在表征水平上，反馈连接显著提升了模型重建模糊或噪声图像的能力。
• SDPC框架成功将神经水平的反馈组织结构与表征水平的抗噪鲁棒性相结合。