Integrating Flexible Normalization into Mid-Level Representations of Deep Convolutional Neural Networks
This work addresses the gap in handling contextual modulation for researchers in computational neuroscience and AI, though it is incremental as it builds on existing normalization approaches.
The authors tackled the problem of contextual effects in deep CNNs for neural prediction by integrating flexible normalization into mid-level representations, resulting in a model that captures spatial dependencies in features like textures and predicts when normalization might be recruited in cortical areas.
Deep convolutional neural networks (CNNs) are becoming increasingly popular models to predict neural responses in visual cortex. However, contextual effects, which are prevalent in neural processing and in perception, are not explicitly handled by current CNNs, including those used for neural prediction. In primary visual cortex, neural responses are modulated by stimuli spatially surrounding the classical receptive field in rich ways. These effects have been modeled with divisive normalization approaches, including flexible models, where spatial normalization is recruited only to the degree responses from center and surround locations are deemed statistically dependent. We propose a flexible normalization model applied to mid-level representations of deep CNNs as a tractable way to study contextual normalization mechanisms in mid-level cortical areas. This approach captures non-trivial spatial dependencies among mid-level features in CNNs, such as those present in textures and other visual stimuli, that arise from tiling high order features, geometrically. We expect that the proposed approach can make predictions about when spatial normalization might be recruited in mid-level cortical areas. We also expect this approach to be useful as part of the CNN toolkit, therefore going beyond more restrictive fixed forms of normalization.