The image generation community may be asking single models to solve too many problems at once, and a broader lesson is that models fail not because of architecture limitations but because of how objectives and training data are provided.

11:46algorithms nowadays. >> Uh yes, you are absolutely right. curriculum now is making impact not only by the way for our papers but there were several other papers I see we are talking about how wonderfully curriculum learning makes let's say multimodel models better one broader lesson from the uh disco paper sometimes the model seems simply miss the right objective in training uh like the things that I mentioned and also um uh it is not

12:18mainly an architecture limit itation but it is how you are providing this objective and training data to the algorithm. The way I would summarize your answer is that yes, there are lots of different attributes that you might want to exert some control over and um you know there is a a kind of a you know a mental or human cost to going after each of these attributes. But from