Abstract: We present an unsupervised visual feature learning algorithm driven by context-based pixel prediction. By analogy with auto-encoders, we propose Context Encoders - a convolutional neural ...
Abstract: A key barrier to mainstream adoption of co-package photonics is a high-yielding and scalable assembly process. The current industry state-of-the-art for edge coupling is to attach fibers ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results