Congratulations to everyone who participated in the Cyber-Physical Anomaly Detection for DER Systems competition!\n The private leaderboard is now final.
\n| Place | \nCompetitor | \nPrivate F2 Score | \n
|---|---|---|
| 1st | \n@alvaroborras | \n0.91742 | \n
| 2nd | \n@geethasivanantham | \n0.91191 | \n
| 3rd | \n@edonisraci | \n0.91190 | \n
The top F2 score of 0.91742 is a strong result on this dataset, well above our internal baselines. The competition\n was tight, with the top 49 teams all scoring above 0.91.
\nPer the competition\n rules, winners are\n required to:
\nSubmit your solution code and methodology — a reproducible pipeline (training + inference) that can\nregenerate your winning submission, along with a description of your approach, preprocessing, model architecture,\nhyperparameters, and computational environment. Code should be shared via a linked repository or Kaggle notebook.\nSee the documentation guidelines for details.
Provide your contact email — so we can coordinate prize delivery and solution verification.
\nDeadline: Friday, April 10, 2026. Please reply to this post or message me directly with your code and contact\ninformation. If a winner does not respond by the deadline, the prize will pass to the next eligible participant.
\nWe will verify that each submission reproduces the claimed score before distributing prizes. Winning solutions\nwill be licensed under Apache 2.0 as stated in the competition rules.
We'd love to see what approaches the community used — whether you placed 1st or 61st. If you're willing, consider\n posting a brief write-up of your methodology in the discussion forum. What features mattered most? What models did\n you try? What didn't work? These posts are one of the best parts of Kaggle competitions and help everyone learn.
\nThis was our first time hosting a competition, and we learned a lot — especially from those of you who helped us\n catch and fix data quality issues early on. Your patience and feedback made this a better competition for\n everyone.
\nThe dataset will remain available under its current license (CC BY-NC-SA 4.0) for academic research and education.\n If you use it in published work, please cite:
\n\n\nJorge Pineda, Jay Johnson. Cyber-Physical Anomaly Detection for DER Systems.\n https://kaggle.com/competitions/cyber-physical-anomaly-detection-for-der-systems. Kaggle, 2026.
\n
Thanks again for contributing to DER security research.
\nJorge Pineda — Competition Host
", "messages": [ { "id": 3437907, "postDate": "2026-04-08T10:37:53.490Z", "content": "is it necessary to share the presentation right now
", "rawMarkdown": "is it necessary to share the presentation right now\n", "votes": 1, "replies": [ { "id": 3438141, "postDate": "2026-04-08T18:06:39.310Z", "content": "Yes or just a github repo, it doesn't need to be public as long as we have access.
", "rawMarkdown": "Yes or just a github repo, it doesn't need to be public as long as we have access." } ] }, { "id": 3436691, "postDate": "2026-04-06T14:30:16.673Z", "content": "Congratulations to everyone who participated in the Cyber-Physical Anomaly Detection for DER Systems competition!\n The private leaderboard is now final.
\n| Place | \nCompetitor | \nPrivate F2 Score | \n
|---|---|---|
| 1st | \n@alvaroborras | \n0.91742 | \n
| 2nd | \n@geethasivanantham | \n0.91191 | \n
| 3rd | \n@edonisraci | \n0.91190 | \n
The top F2 score of 0.91742 is a strong result on this dataset, well above our internal baselines. The competition\n was tight, with the top 49 teams all scoring above 0.91.
\nPer the competition\n rules, winners are\n required to:
\nSubmit your solution code and methodology — a reproducible pipeline (training + inference) that can\nregenerate your winning submission, along with a description of your approach, preprocessing, model architecture,\nhyperparameters, and computational environment. Code should be shared via a linked repository or Kaggle notebook.\nSee the documentation guidelines for details.
Provide your contact email — so we can coordinate prize delivery and solution verification.
\nDeadline: Friday, April 10, 2026. Please reply to this post or message me directly with your code and contact\ninformation. If a winner does not respond by the deadline, the prize will pass to the next eligible participant.
\nWe will verify that each submission reproduces the claimed score before distributing prizes. Winning solutions\nwill be licensed under Apache 2.0 as stated in the competition rules.
We'd love to see what approaches the community used — whether you placed 1st or 61st. If you're willing, consider\n posting a brief write-up of your methodology in the discussion forum. What features mattered most? What models did\n you try? What didn't work? These posts are one of the best parts of Kaggle competitions and help everyone learn.
\nThis was our first time hosting a competition, and we learned a lot — especially from those of you who helped us\n catch and fix data quality issues early on. Your patience and feedback made this a better competition for\n everyone.
\nThe dataset will remain available under its current license (CC BY-NC-SA 4.0) for academic research and education.\n If you use it in published work, please cite:
\n\n\nJorge Pineda, Jay Johnson. Cyber-Physical Anomaly Detection for DER Systems.\n https://kaggle.com/competitions/cyber-physical-anomaly-detection-for-der-systems. Kaggle, 2026.
\n
Thanks again for contributing to DER security research.
\nJorge Pineda — Competition Host
", "rawMarkdown": "Congratulations to everyone who participated in the Cyber-Physical Anomaly Detection for DER Systems competition!\n The private leaderboard is now final.\n\n## Prize Winners\n\n | Place | Competitor | Private F2 Score |\n |-------|-----------|-----------------|\n | 1st | @alvaroborras | 0.91742 |\n | 2nd | @geethasivanantham | 0.91191 |\n | 3rd | @edonisraci | 0.91190 |\n\n The top F2 score of 0.91742 is a strong result on this dataset, well above our internal baselines. The competition\n was tight, with the top 49 teams all scoring above 0.91.\n\n## Next Steps for Prize Winners\n\n Per the [competition\n rules](https://www.kaggle.com/competitions/cyber-physical-anomaly-detection-for-der-systems/rules), winners are\n required to:\n\n 1. **Submit your solution code and methodology** — a reproducible pipeline (training + inference) that can\n regenerate your winning submission, along with a description of your approach, preprocessing, model architecture,\n hyperparameters, and computational environment. Code should be shared via a linked repository or Kaggle notebook.\n See the [documentation guidelines](https://www.kaggle.com/WinningModelDocumentationGuidelines) for details.\n\n 2. **Provide your contact email** — so we can coordinate prize delivery and solution verification.\n\n **Deadline: Friday, April 10, 2026.** Please reply to this post or message me directly with your code and contact\n information. If a winner does not respond by the deadline, the prize will pass to the next eligible participant.\n\n We will verify that each submission reproduces the claimed score before distributing prizes. Winning solutions\n will be licensed under Apache 2.0 as stated in the competition rules.\n\n## Solution Sharing\n\n We'd love to see what approaches the community used — whether you placed 1st or 61st. If you're willing, consider\n posting a brief write-up of your methodology in the discussion forum. What features mattered most? What models did\n you try? What didn't work? These posts are one of the best parts of Kaggle competitions and help everyone learn.\n\n## Thank You\n\n This was our first time hosting a competition, and we learned a lot — especially from those of you who helped us\n catch and fix data quality issues early on. Your patience and feedback made this a better competition for\n everyone.\n\n The dataset will remain available under its current license (CC BY-NC-SA 4.0) for academic research and education.\n If you use it in published work, please cite:\n\n > Jorge Pineda, Jay Johnson. *Cyber-Physical Anomaly Detection for DER Systems.*\n https://kaggle.com/competitions/cyber-physical-anomaly-detection-for-der-systems. Kaggle, 2026.\n\n Thanks again for contributing to DER security research.\n\n Jorge Pineda — Competition Host" }, { "id": 3438504, "postDate": "2026-04-09T09:18:37.530Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 3437907, "author_name": "Geetha Sivanantham", "author_url": "", "post_date": "2026-04-08T10:37:53.490000", "content": "is it necessary to share the presentation right now
", "votes": 1, "replies": [ { "id": 3438141, "author_name": "Jorge Pineda", "author_url": "", "post_date": "2026-04-08T18:06:39.310000", "content": "Yes or just a github repo, it doesn't need to be public as long as we have access.
", "votes": 0, "replies": [] } ] }, { "id": 3438504, "author_name": "", "author_url": "", "post_date": "2026-04-09T09:18:37.530000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "3437907": "is it necessary to share the presentation right now\n", "3436691": "Congratulations to everyone who participated in the Cyber-Physical Anomaly Detection for DER Systems competition!\n The private leaderboard is now final.\n\n## Prize Winners\n\n | Place | Competitor | Private F2 Score |\n |-------|-----------|-----------------|\n | 1st | @alvaroborras | 0.91742 |\n | 2nd | @geethasivanantham | 0.91191 |\n | 3rd | @edonisraci | 0.91190 |\n\n The top F2 score of 0.91742 is a strong result on this dataset, well above our internal baselines. The competition\n was tight, with the top 49 teams all scoring above 0.91.\n\n## Next Steps for Prize Winners\n\n Per the [competition\n rules](https://www.kaggle.com/competitions/cyber-physical-anomaly-detection-for-der-systems/rules), winners are\n required to:\n\n 1. **Submit your solution code and methodology** — a reproducible pipeline (training + inference) that can\n regenerate your winning submission, along with a description of your approach, preprocessing, model architecture,\n hyperparameters, and computational environment. Code should be shared via a linked repository or Kaggle notebook.\n See the [documentation guidelines](https://www.kaggle.com/WinningModelDocumentationGuidelines) for details.\n\n 2. **Provide your contact email** — so we can coordinate prize delivery and solution verification.\n\n **Deadline: Friday, April 10, 2026.** Please reply to this post or message me directly with your code and contact\n information. If a winner does not respond by the deadline, the prize will pass to the next eligible participant.\n\n We will verify that each submission reproduces the claimed score before distributing prizes. Winning solutions\n will be licensed under Apache 2.0 as stated in the competition rules.\n\n## Solution Sharing\n\n We'd love to see what approaches the community used — whether you placed 1st or 61st. If you're willing, consider\n posting a brief write-up of your methodology in the discussion forum. What features mattered most? What models did\n you try? What didn't work? These posts are one of the best parts of Kaggle competitions and help everyone learn.\n\n## Thank You\n\n This was our first time hosting a competition, and we learned a lot — especially from those of you who helped us\n catch and fix data quality issues early on. Your patience and feedback made this a better competition for\n everyone.\n\n The dataset will remain available under its current license (CC BY-NC-SA 4.0) for academic research and education.\n If you use it in published work, please cite:\n\n > Jorge Pineda, Jay Johnson. *Cyber-Physical Anomaly Detection for DER Systems.*\n https://kaggle.com/competitions/cyber-physical-anomaly-detection-for-der-systems. Kaggle, 2026.\n\n Thanks again for contributing to DER security research.\n\n Jorge Pineda — Competition Host", "3438504": "" } }