Question about defining the contrastive loss on zi’s rather than hi’s

[garynlfd]

Hello,

I have a question after reading your paper. You mentioned that it's beneficial to define the contrastive loss on zi’s rather than hi’s, but I'm not sure what's the main reason of this. As you mentioned in section 4, "z is trained to be invariant to data transformation", is this the main reason? Could you give me more evidence about why z is better than h in contrastive loss? I would be grateful if you could give me some hints.

Thanks.