[D] OOD and Spandrels, or What you should know about EBM.
About this article
Energy-based model This article will compare EBMs to multi-layered perceptrons, and addresses a lingering question : Whether or not EBMs are simply an "equivalent reformulation" of traditional MLPs with gradient descent. Given the same training data, and the same parameter count, do EBM simply converge to what would result from a traditional MLP trained by gradient descent? It turns out the answer is no. EBMs differ most sharply from MLP in how they categorize OOD points that are near the bou...
You've been blocked by network security.To continue, log in to your Reddit account or use your developer tokenIf you think you've been blocked by mistake, file a ticket below and we'll look into it.Log in File a ticket