![\"Fig.1\"](\"https://storage.googleapis.com/kaggle-forum-message-attachments/347824/9691/fig1.png\")
\nFig.1 shows some reconstructed tracks exactly the same as GT in event000001001. Some tracks are far from (0,0,z), as we know it is a challenge now. But it is surprised that there are some non-helix tracks could be find.Just find something interesting, my approach finds some tracks beyond my imagination.\n\nFig.1 shows some reconstructed tracks exactly the same as GT in event000001001. Some tracks are far from (0,0,z), as we know it is a challenge now. But it is surprised that there are some non-helix tracks could be find.
The second attached file is the reconstructed result. You may do some analysis as I do now.
", "messages": [ { "id": "347824", "postDate": "06/25/2018 13:45:23", "content": "Just find something interesting, my approach finds some tracks beyond my imagination.\n\nFig.1 shows some reconstructed tracks exactly the same as GT in event000001001. Some tracks are far from (0,0,z), as we know it is a challenge now. But it is surprised that there are some non-helix tracks could be find.
The second attached file is the reconstructed result. You may do some analysis as I do now.
", "rawMarkdown": "Just find something interesting, my approach finds some tracks beyond my imagination.\n![Fig.1][1]\nFig.1 shows some reconstructed tracks exactly the same as GT in event000001001. Some tracks are far from (0,0,z), as we know it is a challenge now. But it is surprised that there are some non-helix tracks could be find.\n\nThe second attached file is the reconstructed result. You may do some analysis as I do now.\n\n\n [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/347824/9691/fig1.png", "votes": null }, { "id": "347839", "postDate": "06/25/2018 14:08:26", "content": "@outrunner, you finally started sharing :D (thumbs up)
", "rawMarkdown": "outrunner, you finally started sharing :D (thumbs up)", "votes": null }, { "id": "347914", "postDate": "06/25/2018 17:00:36", "content": "When you say your approach finds these tracks, do you mean your exploratory analysis is finding tracks that will be difficult to model, or are you saying your model has successfully identified the above tracks?
", "rawMarkdown": "When you say your approach finds these tracks, do you mean your exploratory analysis is finding tracks that will be difficult to model, or are you saying your model has successfully identified the above tracks?", "votes": null }, { "id": "347942", "postDate": "06/25/2018 18:14:01", "content": "Seems you are able to model even the glitches from the simulator. Impressive.
", "rawMarkdown": "Seems you are able to model even the glitches from the simulator. Impressive.", "votes": null }, { "id": "347945", "postDate": "06/25/2018 18:25:04", "content": "\n\nThanks for the file. I briefly analyse your results. I will post more document and diagrams later.\nFor a quick notes:
\n\nif you analyse outrunner results, you will fill that it is very impressive. below shows prediction at r<200 and z<500.
\n\nblack: truth, red: outrunner
\n\n![\"enter](\"https://storage.googleapis.com/kaggle-forum-message-attachments/347946/9695/outrunner1.png\")
Are these really considered 'glitches' from the simulator? My understanding is that the simulation is pretty accurate compared to experimental results. Would we not expect to see more difficult tracks like these in reality?
", "rawMarkdown": "Are these really considered 'glitches' from the simulator? My understanding is that the simulation is pretty accurate compared to experimental results. Would we not expect to see more difficult tracks like these in reality?", "votes": null }, { "id": "348028", "postDate": "06/25/2018 23:03:49", "content": "\n\n\nThanks for the file. I briefly analyse your results. I will post more document and diagrams later.
\n
As you always do the best sharing, and it would be helpful, thank you.
", "rawMarkdown": "> Thanks for the file. I briefly analyse your results. I will post more document and diagrams later.\n\nAs you always do the best sharing, and it would be helpful, thank you.", "votes": null }, { "id": "348029", "postDate": "06/25/2018 23:08:46", "content": "The latter, it is a complex pipeline with endless loop of PostProcess(Merge(Find every single track))). I think most of us lost in it.
", "rawMarkdown": "The latter, it is a complex pipeline with endless loop of PostProcess(Merge(Find every single track))). I think most of us lost in it.", "votes": null }, { "id": "348031", "postDate": "06/25/2018 23:17:09", "content": "\n\n\nSeems you are able to model even the glitches from the simulator. Impressive.
\n
This is beyond my expectation, I thought finding track may be more easily than I know now. And they may be the lucky in extension step.
", "rawMarkdown": "> Seems you are able to model even the glitches from the simulator. Impressive.\n\nThis is beyond my expectation, I thought finding track may be more easily than I know now. And they may be the lucky in extension step.", "votes": null }, { "id": "348048", "postDate": "06/26/2018 01:27:18", "content": "\n\n\nAre these really considered 'glitches' from the simulator?
\n
The organizers admitted that the simulator has glitches, for instance in my post about data quality issues.
", "rawMarkdown": "> Are these really considered 'glitches' from the simulator? \n\nThe organizers admitted that the simulator has glitches, for instance in my post about data quality issues.", "votes": null }, { "id": "348132", "postDate": "06/26/2018 06:34:55", "content": "mistake as attached:
\n\n![\"enter](\"https://storage.googleapis.com/kaggle-forum-message-attachments/348132/9700/misake.png\")
Link to Organizers comments about glitches: https://www.kaggle.com/c/trackml-particle-identification/discussion/57117#335358
", "rawMarkdown": "Link to Organizers comments about glitches: https://www.kaggle.com/c/trackml-particle-identification/discussion/57117#335358", "votes": null }, { "id": "348584", "postDate": "06/27/2018 01:15:09", "content": "These came from the aggressive extension stage (with a bug) in the end. Even though I still miss 0.01~0.02 weights in the free hits. The pipeline only add hits to track right now, I will try to do some \"remove\".
", "rawMarkdown": "These came from the aggressive extension stage (with a bug) in the end. Even though I still miss 0.01~0.02 weights in the free hits. The pipeline only add hits to track right now, I will try to do some \"remove\".", "votes": null }, { "id": "358461", "postDate": "07/18/2018 07:36:15", "content": "There is a zeta and a phi. :-) (tracks reconstructed exactly in event000001001)\n![\"enter](\"https://storage.googleapis.com/kaggle-forum-message-attachments/358461/9920/trackML_9017.png\")
@outrunner, you finally started sharing :D (thumbs up)
", "votes": null, "replies": [] }, { "id": 347914, "author_name": "macfarll", "author_url": "", "post_date": "06/25/2018 17:00:36", "content": "When you say your approach finds these tracks, do you mean your exploratory analysis is finding tracks that will be difficult to model, or are you saying your model has successfully identified the above tracks?
", "votes": null, "replies": [ { "id": 348029, "author_name": "outrunner", "author_url": "", "post_date": "06/25/2018 23:08:46", "content": "The latter, it is a complex pipeline with endless loop of PostProcess(Merge(Find every single track))). I think most of us lost in it.
", "votes": null, "replies": [] } ] }, { "id": 347942, "author_name": "cpmpml", "author_url": "", "post_date": "06/25/2018 18:14:01", "content": "Seems you are able to model even the glitches from the simulator. Impressive.
", "votes": null, "replies": [ { "id": 348017, "author_name": "matthewmasters", "author_url": "", "post_date": "06/25/2018 22:17:29", "content": "Are these really considered 'glitches' from the simulator? My understanding is that the simulation is pretty accurate compared to experimental results. Would we not expect to see more difficult tracks like these in reality?
", "votes": null, "replies": [] }, { "id": 348031, "author_name": "outrunner", "author_url": "", "post_date": "06/25/2018 23:17:09", "content": "\n\n\nSeems you are able to model even the glitches from the simulator. Impressive.
\n
This is beyond my expectation, I thought finding track may be more easily than I know now. And they may be the lucky in extension step.
", "votes": null, "replies": [] }, { "id": 348048, "author_name": "cpmpml", "author_url": "", "post_date": "06/26/2018 01:27:18", "content": "\n\n\nAre these really considered 'glitches' from the simulator?
\n
The organizers admitted that the simulator has glitches, for instance in my post about data quality issues.
", "votes": null, "replies": [] }, { "id": 348194, "author_name": "cpmpml", "author_url": "", "post_date": "06/26/2018 08:25:13", "content": "Link to Organizers comments about glitches: https://www.kaggle.com/c/trackml-particle-identification/discussion/57117#335358
", "votes": null, "replies": [] } ] }, { "id": 347945, "author_name": "hengck23", "author_url": "", "post_date": "06/25/2018 18:25:04", "content": "\n\nThanks for the file. I briefly analyse your results. I will post more document and diagrams later.\nFor a quick notes:
\n\n\n\n\nThanks for the file. I briefly analyse your results. I will post more document and diagrams later.
\n
As you always do the best sharing, and it would be helpful, thank you.
", "votes": null, "replies": [] } ] }, { "id": 347946, "author_name": "hengck23", "author_url": "", "post_date": "06/25/2018 18:27:47", "content": "if you analyse outrunner results, you will fill that it is very impressive. below shows prediction at r<200 and z<500.
\n\nblack: truth, red: outrunner
\n\n
mistake as attached:
\n\n
These came from the aggressive extension stage (with a bug) in the end. Even though I still miss 0.01~0.02 weights in the free hits. The pipeline only add hits to track right now, I will try to do some \"remove\".
", "votes": null, "replies": [] } ] }, { "id": 358461, "author_name": "outrunner", "author_url": "", "post_date": "07/18/2018 07:36:15", "content": "There is a zeta and a phi. :-) (tracks reconstructed exactly in event000001001)\n