February 13, 2026 Master Thesis Open Access

Baptist, Frank Michael

JSON Export

{
  "conceptrecid": "22416", 
  "created": "2026-03-24T15:05:38.972038+00:00", 
  "files": [
    {
      "bucket": "a5f98984-530f-47fc-96fa-69a784683473", 
      "checksum": "md5:e79ef60462966d2deef2030cca66da26", 
      "key": "Frank_Masterthesis-3.pdf", 
      "links": {
        "self": "https://publish.etp.kit.edu/api/files/a5f98984-530f-47fc-96fa-69a784683473/Frank_Masterthesis-3.pdf"
      }, 
      "size": 11387675, 
      "type": "pdf"
    }
  ], 
  "id": 22417, 
  "links": {
    "bucket": "https://publish.etp.kit.edu/api/files/a5f98984-530f-47fc-96fa-69a784683473", 
    "html": "https://publish.etp.kit.edu/record/22417", 
    "latest": "https://publish.etp.kit.edu/api/records/22417", 
    "latest_html": "https://publish.etp.kit.edu/record/22417"
  }, 
  "metadata": {
    "access_right": "open", 
    "access_right_category": "success", 
    "communities": [
      {
        "id": "belle2"
      }, 
      {
        "id": "etp"
      }, 
      {
        "id": "hardware"
      }
    ], 
    "contributors": [], 
    "creators": [
      {
        "affiliation": "KIT/ETP", 
        "name": "Baptist, Frank Michael"
      }
    ], 
    "description": "<p>At the Belle II experiment, collisions occur at a rate higher than 200 MHz. Due to bandwidth and storage limitations, storing this amount of data for offline analysis is not feasible. That is why the experiment relies on a two-level trigger system as a preselection before passing the events to subsequent software-based processing and analysis. The first level, known as the Level 1 Trigger (L1 Trigger), is a hardware-based system that makes decisions within a few microseconds. The buffered data from the detector can only be stored for a limited time, which imposes strict latency requirements on the L1 Trigger.</p>\n\n<p>One approach to enhance the L1 Trigger is the development of the GNN-ETM. It utilizes dynamic graph building based on GravNet and can predict a varying number of clusters per event using Object Condensation. The GNN-ETM is implemented on field-programmable gate array (FPGA) hardware, which allows for low-latency inference and is suitable for deployment on the L1 Trigger. The current parasitic implementation of the GNN-ETM in the L1 Trigger has shown promising results in terms of precision and latency. Nevertheless, further improvements, especially in terms of latency, are necessary to develop a potential replacement for the clustering algorithm currently deployed on the ECL L1 Trigger.</p>\n\n<p>In this thesis, I present the optimization of the GNN-ETM for reduced latency, including modifications to the model architecture and training procedure.</p>", 
    "keywords": [
      "GNN-ETM", 
      "GNN", 
      "Belle II", 
      "Trigger", 
      "GravNet", 
      "Object Condensation"
    ], 
    "language": "eng", 
    "publication_date": "2026-02-13", 
    "relations": {
      "version": [
        {
          "count": 1, 
          "index": 0, 
          "is_last": true, 
          "last_child": {
            "pid_type": "recid", 
            "pid_value": "22417"
          }, 
          "parent": {
            "pid_type": "recid", 
            "pid_value": "22416"
          }
        }
      ]
    }, 
    "resource_type": {
      "subtype": "master-thesis", 
      "title": "Master Thesis", 
      "type": "thesis"
    }, 
    "thesis": {
      "report_number": "ETP-KA/2026-03", 
      "supervisors": [
        {
          "affiliation": "KIT/ETP", 
          "name": "Ferber, Torben"
        }, 
        {
          "affiliation": "KIT/ETP", 
          "name": "Klute, Markus"
        }, 
        {
          "affiliation": "KIT/ETP", 
          "name": "Haide, Isabel"
        }
      ]
    }, 
    "title": "Training and Optimization of a Graph Neural Network for Deployment on FPGA Hardware in the Belle II Level 1 Trigger"
  }, 
  "owners": [
    180
  ], 
  "revision": 4, 
  "stats": {}, 
  "updated": "2026-03-24T15:05:44.096925+00:00"
}