Daniel Humphreys

My research:

My primary analysis is a search for Higgs light resonances. Specifically I am conducting a search for the process $S\rightarrow Z_d Z_d \rightarrow \ell\ell\nu\nu$, utilizing parameterized neural networks. I am also working on a GNN implemented level 0 trigger for the ATLAS muon spectrometer.

My expertise is:

• General data analysis.
  • Familiar with multiple neural network architectures
  • Processing raw binary data
  • Proficient in object oriented programming with Python and C++
• Being thrown on to new projects I know nothing about.
• Arguing with chatgpt

A problem I’m grappling with:

How to transition a GNN from being a hit classifier to an edge/track classifier

I’ve got my eyes on:

Learning about heterogeneous processing. I find GPU’s to be really fascinating and have potential for very large performance improvements.

I want to know more about:

I’d like to dig deeper into Graphical Neural Networks and Transformers. It would be really cool to build my own LLM.