Andreas Michael Hermansen

icon company University of Geneva icon location PhD student at University of Geneva, working on Machine Learning applied to the ATLAS experiment.

Results 7 comments of Andreas Michael Hermansen

Deprecation List for GraphNeT 2.0

- [ ] - `save_model_config` and `save_dataset_config` decorators. See #591

Deprecation List for GraphNeT 2.0

We should probably have a look at the functions mentioned in #467 and decide which of those are deprecated and mark them as such, before potentially removing them.

Add example directory and example scripts of using a pretrained model

> I was looking for how to apply it to a GraphNeT dataset, I know it shouldn't be too difficult to do, I'm just rather unsure about what can be...

Use unsupervised learning to learn representations in data and Monte Carlo

This is the current layout of the MAE that I've been working. The general MAE architecture is based of [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) ![GraphNeT_BEiT_Style_MAE drawio](https://github.com/graphnet-team/graphnet/assets/97125645/cc13d0bf-8b7b-4d04-9dd3-87470bbc6fb2) The aggregation loss...

Use unsupervised learning to learn representations in data and Monte Carlo

> Hey @AMHermansen I've now taken a deeper look at this, and I must admit I still struggle to follow your diagram. I _think_ that your diagram depicts the usual...

Use unsupervised learning to learn representations in data and Monte Carlo

> @AMHermansen thanks for the details! Before I respond in full, could you help clear up one last thing: > > > To get the summary loss a new module...

Use unsupervised learning to learn representations in data and Monte Carlo

Sorry for the wait. > Alright @AMHermansen - sorry for the delay on this. I now have some pseudo-code ready. > > I arrived at this pseudo-code by starting out...