cellular structure are same.(bio pic)
hierarchical structure.(need a pic)and each region is performing the same set of processes on the input data.
[SDR] sparse distributed representations(0 or 1)
Input: 1. Motor commands 2. sensory input
Encoder: takes a datatype and converts it into a sparse distributed representations.
Temporal means that systems learn continuously, every time it receives input it is attemptiing to predict what is going to happen next.
- Sparse Distributed Representation(SDR)
Terms: 1. n=Bit array length 2. w=Bits of array 3. sparsiity 4. Dense Bit Array Capacity=2**of bits
Similarity can be represented by overlap/union? of SDR.
- Overlap Sets and Subsampling
- Scalar Encoder: consecutive one
- Random Distributed Scalar Encoder: random one
- Input Space& Connections
- Spactial Pooler: maintain a fixed sparsity & maintain a overlap properties.