Your best bet for ensuring that you always get a fresh cake is to plan ahead Tampa Bay Buccaneers Selmon Brady Brooks Champion signatures shirt and order exactly what you want. The more you leave to chance, the less control you have. Make the cake yourself. I know we are all getting really used to living in a world of instant gratification, but honestly, it is not that hard. It’s not like curing cancer. All you have to do is be able to follow simple instructions and have a few basic ingredients on hand. I won’t insult your intelligence by assuming you don’t know how to Google a cake recipe. If we only use the reinforcement signal to guide training, then I agree with Yann LeCun that it is the cherry on the cake. Even worse: when using a global reinforcement signal that is not a known differentiable function of the representations (which is typically the case), there is a serious scaling problem in terms of the number of hidden units (or action dimensions) that can be trained with respect to that signal.
The number of examples, random samples, or trials of actions may have to grow at least linearly with the number of units in order to provide credit assignment of quality comparable to that obtained with back-propagation. If the action space is large, this is problematic. However, as Demis Hassabis said when Yann talked about the cake and cherry analogy, we should *also* do unsupervised learning, along with reinforcement learning. Then it becomes more credible that it can work on a large scale. It’s a pretty easy reasoning question but if they do not proceed carefully then such questions can be turned into a big mess, even a silly mistake could ruin the entire labor. At some point, an immature individual who didn’t want to follow the actual tradition – the bride and groom feeding each other a small first bite of the cake – mashed the piece of cake into his bride’s face. The practice spread, and after a while, brides started to retaliate and smash cake back.