https://abaj.ai/tags/parameter-sharing/Recent content in Parameter-Sharing on Aayush Bajaj's Augmenting InfrastructureHugoen© 2026 Aayush BajajThu, 09 Jul 2026 21:02:20 +1000https://abaj.ai/wiki/ml/dl/cnn/Thu, 09 Jul 2026 21:02:56 +1000https://abaj.ai/wiki/ml/dl/cnn/<p>a convolutional neural network is a <a href="https://abaj.ai/wiki/ml/dl/feedforward/" >feedforward network</a> with its linear layers put on a diet: instead of every unit seeing every input, each unit sees a small local window, and every window is processed by the <em>same</em> small set of weights.<span class="margin-note" data-note="the licence for this diet is an assumption about images: nearby pixels are related, faraway pixels mostly are not, and a feature worth detecting is worth detecting everywhere"> <span class="margin-note-indicator">𐃏</span> </span> this page builds the operation from its definition, gets the geometry formulas straight, walks the canonical architectures, and ends with a convolution written in loops and checked against scipy (Goodfellow, Ian, 2016).</p>