AI RESEARCH
Parameter-Efficient Architectural Modifications for Translation-Invariant CNNs
arXiv CS.CV
•
ArXi:2604.27870v1 Announce Type: new Convolutional Neural Networks (CNNs) are widely assumed to be translation-invariant, yet standard architectures exhibit a startling fragility: even a single-pixel shift can drastically degrade performance due to their reliance on spatially dependent fully connected layers. In this work, we resolve this vulnerability by proposing a lightweight 'Online Architecture' strategy. By strategically inserting Global Average Pooling (GAP) layers at various network depths, we effectively decouple feature recognition from spatial location.