On Energy-Based Models with Overparametrized Shallow Neural Networks

          Carles Domingo-Enrich 1 Alberto Bietti 2 Eric Vanden-Eijnden 1 Joan Bruna 1 2

             Abstract                  tional to exp{f (x)}. Such energy-based models (EBMs)
                                   originate in statistical physics (Ruelle, 1969), and have be-
    Energy-based models (EBMs) are a simple yet          come a fundamental modeling tool in statistics and machine
    powerful framework for generat ...