hi,
there is some paper about Spatial pooling and the authors said it is good sparse representation"such as The HTM Spatial Pooler -A Neocortical Algorithm for Online Sparse Distributed Coding",I want to know is there any body that compares this representation by other sparse representation? is there any paper about this?how much this method is better?
thanks a lot
Not that I know of. Do you have any examples of other semantic sparse representation formats you wish to compare? There are also 3 videos about SDRs and how they work that might enlighten you if you are just getting started with this topic:
No,l Don compare sp by any other sparse representation,but I know there are many different methods for sparse representation, I like to know why sp is better than other way,why you use this methode?
In machine learning or just in general? Like random drop-out?
Because that’s how the brain seems to do it. Have you seen either of these videos: Spatial Pooling: Input Space & Connections or Spatial Pooling: Learning(Episode 8)? Have you read the spatial pooler chapter of BAMI?
Yes, In machine learning.
You know I really like to see the good performance of sp,for example I like to get mnist dataset to sp and one other sparse methode and see that sp is better.
But I have no idea about other methods that can generate sparse data like sp, and I don’t have implementation
I I will be very thankful if you help me
I’m not sure I can help. We only study how the brain creates sparsity, like the SP. We are not too interested in other methods.
1 Like