This week's assignments will guide you through the following topics:
- Deep learning based on low-level features
- Embedding the inputs
- Choosing a neural network architecture
Please skim the following:
- Jet images: Ref. {cite:p}
deOliveira:2015xxd
- Particle feature lists: Ref. {cite:p}
deepjet
- Sets: Ref. {cite:p}
Komiske:2018cqr
- Graphs/point clouds: Ref. {cite:p}
Moreno:2019bmu,Moreno:2019neq,Qu:2019gqs
Complete the following tasks:
- Run through the notebook 05-deep-learning
- Create a baseline deep learning model with PF candidate, track, and/or secondary vertex features. The choice of input embedding and classifier is up to you. (Note: you do not need to implement the interaction network here).
Answer the following questions on Canvas:
- What are the different ways to encode low-level particle information that you read about to input to a deep learning model?
- What are the different types of neural network architectures that are commonly used for each?