Large-Margin Contrastive Learning with Distance Polarization Regularizer

          Shuo Chen 1 Gang Niu 1 Chen Gong 2 Jun Li 2 Jian Yang 2 Masashi Sugiyama 1 3

               Abstract                      Intra/Inter-Cluster Distance      Enlargement    Polarization
    Contrastive learning (CL) pretrains models in a
    pairwise manner, where given a data point, other
    data points are all regarded as dissimilar, includ-
    ing some that are semantically similar. The issue
    has been add ...