ProofAtlas

Jensen–Shannon distance

TODO. Will cover:

  • Stationary-neighbourhood distributions pu,pvp_u, p_v.
  • JS(pu,pv)=12KL(pum)+12KL(pvm)\mathrm{JS}(p_u, p_v) = \tfrac{1}{2}\, \mathrm{KL}(p_u \,\|\, m) + \tfrac{1}{2}\, \mathrm{KL}(p_v \,\|\, m) with m=12(pu+pv)m = \tfrac{1}{2}(p_u + p_v).
  • JS\sqrt{\mathrm{JS}} is a metric (Endres–Schindelin).
  • IsHypergraphMetric instance status.