Sdr (SDR v0.1.0) View Source

SDR is an Elixir library for Sparse Distributed Representations

Link to this section Summary

Functions

Capacity of a SDR.

Hash encoder.

Overlap of two SDRs.

Overlap of two random similar SDRs given their number of bits n and the number of on bits w.

Linear encoder.

Sparsity of a SDR.

Union of two SDRs.

Union of two random similar SDRs given their number of bits n and the number of on bits w.

Link to this section Functions

Capacity of a SDR.

Examples

      iex(1)> Sdr.capacity(2048, 6)
      101733385755251712
Link to this function

hash(min, max, lc, w, v)

View Source

Hash encoder.

Examples

      iex(1)> Sdr.hash(0, 100, 1, 3, 72)
      #MapSet<["32BB90E8976AAB5298D5DA10FE66F21D", "AD61AB143223EFBC24C7D2583BE69251", "D2DDEA18F00665CE8623E36BD4E3C7C5"]>

Use MapSet.size/1 to get the length of the MapSet. Use MapSet.to_list/1 to convert it to a list.

Overlap of two SDRs.

Examples

      iex(1)> Sdr.overlap(MapSet.new([1, 2]), MapSet.new([2, 3]))
      #MapSet<[2]>

Use MapSet.size/1 to get the length of the MapSet. Use MapSet.to_list/1 to convert it to a list.

Overlap of two random similar SDRs given their number of bits n and the number of on bits w.

Link to this function

simple(min, max, lc, w, v)

View Source

Linear encoder.

Examples

      iex(1)> Sdr.simple(0, 100, 1, 21, 72)
      #MapSet<[72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92]>
      iex(1)> Sdr.simple(0, 100, 1, 21, 73)
      #MapSet<[73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93]>

Use MapSet.size/1 to get the length of the MapSet. Use MapSet.to_list/1 to convert it to a list.

Sparsity of a SDR.

Examples

      iex(1)> Sdr.sparsity(2048, 6)
      0.0029296875

Union of two SDRs.

Examples

      iex(1)> Sdr.union(MapSet.new([1, 2]), MapSet.new([2, 3]))
      #MapSet<[1, 2, 3]>

Use MapSet.size/1 to get the length of the MapSet. Use MapSet.to_list/1 to convert it to a list.

Union of two random similar SDRs given their number of bits n and the number of on bits w.