High-dimensional and Permutation Invariant Anomaly Detection
This addresses the problem of detecting new physics processes in particle physics, where existing methods struggle with high-dimensional, permutation-invariant data, representing a novel method for a known bottleneck.
The paper tackles the challenge of anomaly detection in high-dimensional particle physics data by introducing a permutation-invariant density estimator based on diffusion models, which handles variable-length inputs and achieves effective identification of anomalous jets with low likelihood under background-only conditions.
Methods for anomaly detection of new physics processes are often limited to low-dimensional spaces due to the difficulty of learning high-dimensional probability densities. Particularly at the constituent level, incorporating desirable properties such as permutation invariance and variable-length inputs becomes difficult within popular density estimation methods. In this work, we introduce a permutation-invariant density estimator for particle physics data based on diffusion models, specifically designed to handle variable-length inputs. We demonstrate the efficacy of our methodology by utilizing the learned density as a permutation-invariant anomaly detection score, effectively identifying jets with low likelihood under the background-only hypothesis. To validate our density estimation method, we investigate the ratio of learned densities and compare to those obtained by a supervised classification algorithm.