Hackathon sessions will take place on Thursday and Friday September 3-4. Please fill out this participation form by August 19 to confirm participation in the hackathon, indicate your project interests, and to receive information on computing resources for use in the hackathon. Hackathon teams will be assigned at the beginning of the conference.
Participation modality
The hackathon will be fully in person, with no remote participation option.
Computing resources
Computing resources for the hackathon including GPU resources will be provided by the National Research Platform.
Timetable
The hackathon will take place during 4 formally organized session blocks on Thursday and Friday September 3-4) (although teams are welcome to work together outside of officially programmed sessions). Teams of up to 5 members will be assigned to challenges based on interest and availability at the start of the conference. The hackathon will conclude during the 4th session on Friday from 11:00am - 12:30 pm, where teams will give short summary presentations on their work and judges will choose winning teams and award prizes based on their project and presentation. Separate prizes will be awarded for both hackathon tracks.
Tracks
Participants have a choice of joining a pre-organized or crowdsourced hackathon track:
- Hackathon Track 1 (pre-organized challenges): Challenges based on the COLLIDE2V (COmprehensive Lhc coLlision Dataset for Foundation model dEvelopment) dataset and challenges. These are pre-defined challenges based on this open HEP dataset and provided by the COLLIDE2V organizers, including specific tasks, evaluation metric, and baseline. Teams will be evaluated based on the given evaluation metrics and their results compared to baseline if applicable, and their final presentation. Possible challenge topics in this track are the following (pending sufficient interest in the participation form):
- Challenge 1: Di-Higgs (HH→4b vs. QCD) discrimination for trigger systems
- Challenge 2: Anomaly Detection in the trigger environment
- Challenge 3: Real-time Multiclass Jet Tagging
- Challenge 4: Real-time Missing Transverse Energy Regression
- Challenge 5: Collider Data Foundation Model Pre-training and Online Adaptation
- Challenge 6: Robust Tagging and Self-adapting Trigger Systems Under Changing Detector Conditions
- Hackathon Track 2 (crowdsourced challenges): As an alternative to the COLLIDE2V pre-organized challenge track, you have the option to propose your own hackathon challenge. This can be any challenge that is fastML related from any field that uses an open or accessible dataset defines clear goals and evaluation metrics. Projects can be proposed via the participation form. When teams are assigned at the start of the conference they will have the option to join your proposed project as you describe in the form. If your project is selected as a challenge, you will be assigned to that project team for the hackathon. Proposed projects will be asked the following questions:
-
Title of your challenge
-
Describe your challenge: what is the task you are trying to accomplish?
-
How does this challenge relate to FastML?
-
What dataset will you use for this challenge? Is the dataset open access? Please provide a link to the dataset if possible.
-
How should the challenge be evaluated? (AUC, accuracy/efficiency, resource utilization, interpretability/novelty...)
-
(optional) is there an existing baseline or code people can use as a starting point? Please provide links if applicable.
-
Missing sessions
If you are participating in the hackathon, you will be able to participate in up to one birds-of-a-feather section of your choosing during one of the hackathon session blocks. Please plan to attend all other hackathon sections on Thursday and Friday in order to ensure fair and balanced teams. If you are not participating in the hackathon, you may participate in any or all birds-of-a-feather discussion sessions according to your preference.