- Type
- Research
- Focus
- Transformers · Sensor data
- Context
- Undergraduate thesis · Sep 2023–Jun 2024
My contribution
Developed a deep neural network for neutrino direction reconstruction, integrating GPT-style attention blocks into a Transformer for long-sequence modelling.
Research context
This undergraduate thesis ran from September 2023 to June 2024 at Shanghai Institute of Technology. It addressed long-sequence modelling of complex sensor observations to infer neutrino movement directions.
Recognition
The research approach and final paper received the institute's Outstanding Undergraduate Thesis recognition.