This paper discusses model-agnostic searches for new physics at the Large Hadron Collider using anomaly-detection techniques for the identification of event signatures that deviate from the Standard Model (SM). We investigate anomaly detection in the context of a machine-learning approach based on autoencoders. The analysis uses Monte Carlo simulations for the SM background and several selected exotic models. We also investigate the input space for the event-based anomaly detection and illustrate the shapes of invariant masses in the outlier region which will be used to perform searches for resonant phenomena beyond the SM. Challenges and conceptual limitations of this approach are discussed.

中文翻译：

---

用于搜索新物理的基于事件的异常检测

本文讨论了使用异常检测技术在大型强子对撞机上寻找新物理的模型无关搜索，以识别偏离标准模型 (SM) 的事件特征。我们在基于自动编码器的机器学习方法的背景下研究异常检测。该分析对 SM 背景和几个选定的外来模型使用 Monte Carlo 模拟。我们还研究了基于事件的异常检测的输入空间，并说明了异常区域中不变质量的形状，这些形状将用于搜索 SM 之外的共振现象。讨论了这种方法的挑战和概念限制。