Hashing Function for RDSE

Trying to write RDSE from scratch in Python 3, can’t figure out how to implement the hashing function. From what I read, you need a uniform hash function, is that right? But how do you map the hash to the matrix?


Not sure if this helps, but I have written one in JavaScript. Most is original code, although some of it is almost identical to NuPIC. But my comments are a little different, so maybe they’ll help you?


Many months later, decided to try figure out how to implement RDSE again. Clearly didn’t understand anything about RDSE. For anybody else trying to create their own RDSE from scratch, go read this thread RDSE implementation is wasting buckets? Basically do not think “encoder” as in traditional encoder, it’s all stateful here, load up on memories.

1 Like