EveNet: Towards a Generalist Event Transformer for Unified Understanding and Generation of Collider Data

Ting-Hsiang Hsu², Qibin Liu³, Yuan-Tang Chou¹, Wei-Po Wang², Yue Xu¹, Haoran Zhao¹, Bai-Hong Zhou³, Yi Ren Wu^2*, Shu Li³, Benjamin Nachman⁴, Shih-Chieh Hsu¹, Vinicius Massami Mikuni⁵, Yulei Zhang¹

¹Department of Physics, University of Washington, Seattle, USA
²Department of Physics, National Taiwan University, Taipei, Taiwan
³Department of Physics, Shanghai Jiao Tong University, Shanghai, China
⁴Department of Physics, Stanford University, California, USA
⁵Physics Division, Lawrence Berkeley National Laboratory, California, USA
⁶National Energy Research Scientific Computing Center, California, USA

* Presenter:Yi Ren Wu, email:b11202011@ntu.edu.tw

With the increasing size of the machine learning (ML) model and vast datasets, the foundation model has transformed how we apply ML to solve real-world problems. Multimodal language models like chatGPT and Llama have expanded their capability to specialized tasks with common pre-train. Similarly, in high-energy physics (HEP), common tasks in the analysis face recurring challenges that demand scalable, data-driven solutions. In this talk, we present a foundation model for high-energy physics. Our model leverages extensive simulated datasets in pre-training to address common tasks across analyses, offering a unified starting point for specialized applications. We demonstrate the benefit of using such a pre-train model in improving search sensitivity, anomaly detection, event reconstruction, feature generation, and beyond. By harnessing the power of pre-trained models, we could push the boundaries of discovery with greater efficiency and insight.

Keywords: Machine Learning , Collider Experiments, Foundation Model