Ongoing and Past Work - atlas-calo-ml/Organization GitHub Wiki
Here we should keep a growing list of the things people are trying. The idea is to keep people up to date on what everyone is working on, and indicate what we have explored in the past (with what results), and what we haven't tried. Feel free to edit! This is still a work in progress, and will improve in the next weeks.
| Idea | Person | Result? | Link |
|---|---|---|---|
| Conversion script for new format to images | Albert and Jan and Matthew S | ||
| Testing ResNet with full granularity | Jan | Similar to simple CNN | |
| Testing best CNN with full granularity | Laura | ? | |
| Event displays | Jonathan, Russell, Albert | ||
| Graphs: kNN | Sanmay, Ben, Ming | ||
| Graphs: complicated tensors | Sanmay | ||
| Graphs: simple geometric | Sanmay | ||
| Graphs: neighbors from geometry | Piyush | ||
| Testing performance with jets | Jan | Promising-- understanding strange response binning | |
| Testing performance with taus | Christina | ||
| Testing performance with rho samples | Albert | ongoing | |
| Testing performance with delta samples | Kaela | ongoing | |
| Studying overlapping particle properties in rho and delta | Albert, Kaela | ongoing | |
| Learning resolutions with quantile learning | Albert | Works! | |
| Learning resolutions with MDN | UW team | Works! | |
| Absolute x/y location of tracks? | ? | ||
| Dead material and OOC | Piyush and Dewen | ||
| Baseline classifier and regression | Max, Piyush, Jan | Works! Needs updating to latest samples | |
| Deep sets/ PFN for regression/classification | Russell, Dilia, Kaela, Max | ||
| Deep sets with tracking information | Russell | ||
| Add PFO to outputs for comparisons | ?? |