Finding a good architecture and parameters for convolutional neural networks

I found that the topology and (hyper)parameters of convolutional neural networks can have a dramatic effect on prediction performance, however I haven't found a better way than trial-and-error to find good archictectures. Has anyone find a more insightful way to go about this process?

Created by Stackd

Your web browser must have JavaScript enabled in order for this application to display correctly.
If you are an automated web crawler from a search engine, follow this AJAX application crawl link

I found that the topology and (hyper)parameters of convolutional neural networks can have a dramatic effect on prediction performance, however I haven't found a better way than trial-and-error to find good archictectures. Has anyone find a more insightful way to go about this process?

Drop files to upload

Finding a good architecture and parameters for convolutional neural networks page is loading…