Classifying anomalies through outer density estimation

We propose a new model-agnostic search strategy for physics beyond the standard model (BSM) at LHC, based on novel application of neural density estimation to anomaly detection. Our approach, which we call classifying anomalies through outer (cathode), assumes BSM signal is localized in region (defined e.g., using invariant mass). By training conditional estimator collection additional features outside region, interpolating it into and sampling from it, produce events that follow background model. can then train classifier distinguish data sampled model, thereby approaching optimal detector. Using LHC Olympics R dataset, demonstrate cathode nearly saturates best possible performance, significantly outperforms other approaches aim enhance bump hunt (cwola hunting anode). Finally, very robust against correlations between maintains performance even this more challenging setting.