As I'm sure you're aware deep learning work better is the most hot and highly funded thing right now... there is nothing wrong with making sure you get the credit for a new technique, and it in no way precludes widespread free sharing of the idea.
I am an academic researcher (not in this field) and everything I do is 100% free and public, but I still either patent or publish it, because getting the credit is what allows me to keep doing what I love. I actually hate that part, and would rather just do the work, but I already push that line too far sometimes, and risk losing my funding aka "publish or perish."
I am an academic researcher (not in this field) and everything I do is 100% free and public, but I still either patent or publish it, because getting the credit is what allows me to keep doing what I love. I actually hate that part, and would rather just do the work, but I already push that line too far sometimes, and risk losing my funding aka "publish or perish."