{
  "id": 18346,
  "title": "[Answers] Q&A with Principle Investigators Michael Hansen, Ph.D., and Dr. Andrew Arai of NIH/NHLBI",
  "url": "/competitions/second-annual-data-science-bowl/discussion/18346",
  "author_name": "",
  "post_date": "2016-01-11T16:14:13.603Z",
  "votes": 6,
  "comment_count": 1,
  "views": 1465,
  "content": "<p><a href=\"https://youtu.be/wXUt_vjrF0c\">https://youtu.be/wXUt_vjrF0c</a></p>\n\n<p>Good Monday morning kagglers! <a href=\"https://www.kaggle.com/c/second-annual-data-science-bowl/forums/t/18198/q-a-with-principle-investigators-michael-hansen-ph-d-and-dr-andrew-arai\">As promised</a>, here's a video capture of my offline Q&amp;A with Principle Investigators Michael Hansen, Ph.D., and Dr. Andrew Arai. Thanks again for their time. </p>\n\n<p>Here's a rough sketch of timestamps for the questions and answers:</p>\n\n<p>0:00 - Intros</p>\n\n<p>1:15 - Question 1 -What was the motivation for choosing to evaluate the models based on predictions of systolic and diastolic volumes rather than on ejection fraction (EF)?  </p>\n\n<p>3:43 - Question 2 -What scores would correspond to a model that would be useful in a clinical setting? </p>\n\n<p>11:13  - Question 3 - For each case how many experts segmentation did you use to compute the ground truth volume? </p>\n\n<p>15:12 - Question 4 - How many different experts contributed to the challenge? Are the validation and test sets created by the same experts?</p>\n\n<p>15:24 - Question 5 - How were the minimum and maximum volumes chosen? [I think he means how did you choose systole and diastole] Is it the min/max of the volume along the whole sequence or do you restrict the frame used?</p>\n\n<p>Let me know whether you find this helpful and I should plan on another one, subject to the PIs' availability.</p>",
  "messages": [
    {
      "id": "104307",
      "postDate": "01/11/2016 16:14:13",
      "content": "<p><a href=\"https://youtu.be/wXUt_vjrF0c\">https://youtu.be/wXUt_vjrF0c</a></p>\n\n<p>Good Monday morning kagglers! <a href=\"https://www.kaggle.com/c/second-annual-data-science-bowl/forums/t/18198/q-a-with-principle-investigators-michael-hansen-ph-d-and-dr-andrew-arai\">As promised</a>, here's a video capture of my offline Q&amp;A with Principle Investigators Michael Hansen, Ph.D., and Dr. Andrew Arai. Thanks again for their time. </p>\n\n<p>Here's a rough sketch of timestamps for the questions and answers:</p>\n\n<p>0:00 - Intros</p>\n\n<p>1:15 - Question 1 -What was the motivation for choosing to evaluate the models based on predictions of systolic and diastolic volumes rather than on ejection fraction (EF)?  </p>\n\n<p>3:43 - Question 2 -What scores would correspond to a model that would be useful in a clinical setting? </p>\n\n<p>11:13  - Question 3 - For each case how many experts segmentation did you use to compute the ground truth volume? </p>\n\n<p>15:12 - Question 4 - How many different experts contributed to the challenge? Are the validation and test sets created by the same experts?</p>\n\n<p>15:24 - Question 5 - How were the minimum and maximum volumes chosen? [I think he means how did you choose systole and diastole] Is it the min/max of the volume along the whole sequence or do you restrict the frame used?</p>\n\n<p>Let me know whether you find this helpful and I should plan on another one, subject to the PIs' availability.</p>",
      "rawMarkdown": "https://youtu.be/wXUt_vjrF0c\r\n\r\nGood Monday morning kagglers! [As promised][1], here's a video capture of my offline Q&A with Principle Investigators Michael Hansen, Ph.D., and Dr. Andrew Arai. Thanks again for their time. \r\n\r\nHere's a rough sketch of timestamps for the questions and answers:\r\n\r\n0:00 - Intros\r\n\r\n1:15 - Question 1 -What was the motivation for choosing to evaluate the models based on predictions of systolic and diastolic volumes rather than on ejection fraction (EF)?  \r\n\r\n3:43 - Question 2 -What scores would correspond to a model that would be useful in a clinical setting? \r\n\r\n11:13  - Question 3 - For each case how many experts segmentation did you use to compute the ground truth volume? \r\n\r\n15:12 - Question 4 - How many different experts contributed to the challenge? Are the validation and test sets created by the same experts?\r\n\r\n15:24 - Question 5 - How were the minimum and maximum volumes chosen? [I think he means how did you choose systole and diastole] Is it the min/max of the volume along the whole sequence or do you restrict the frame used?\r\n\r\nLet me know whether you find this helpful and I should plan on another one, subject to the PIs' availability.\r\n\r\n\r\n  [1]: https://www.kaggle.com/c/second-annual-data-science-bowl/forums/t/18198/q-a-with-principle-investigators-michael-hansen-ph-d-and-dr-andrew-arai",
      "votes": null
    },
    {
      "id": "104315",
      "postDate": "01/11/2016 17:30:17",
      "content": "<p>@Shannon, Thanks for putting this together.  And thanks to Dr. Hansen and Dr. Arai for their time.  I found the discussion informative and it gave me some new ideas to try with my model. </p>",
      "rawMarkdown": "Shannon, Thanks for putting this together.  And thanks to Dr. Hansen and Dr. Arai for their time.  I found the discussion informative and it gave me some new ideas to try with my model.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 104315,
      "author_name": "bitsofbits",
      "author_url": "",
      "post_date": "01/11/2016 17:30:17",
      "content": "<p>@Shannon, Thanks for putting this together.  And thanks to Dr. Hansen and Dr. Arai for their time.  I found the discussion informative and it gave me some new ideas to try with my model. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "104307": "https://youtu.be/wXUt_vjrF0c\r\n\r\nGood Monday morning kagglers! [As promised][1], here's a video capture of my offline Q&A with Principle Investigators Michael Hansen, Ph.D., and Dr. Andrew Arai. Thanks again for their time. \r\n\r\nHere's a rough sketch of timestamps for the questions and answers:\r\n\r\n0:00 - Intros\r\n\r\n1:15 - Question 1 -What was the motivation for choosing to evaluate the models based on predictions of systolic and diastolic volumes rather than on ejection fraction (EF)?  \r\n\r\n3:43 - Question 2 -What scores would correspond to a model that would be useful in a clinical setting? \r\n\r\n11:13  - Question 3 - For each case how many experts segmentation did you use to compute the ground truth volume? \r\n\r\n15:12 - Question 4 - How many different experts contributed to the challenge? Are the validation and test sets created by the same experts?\r\n\r\n15:24 - Question 5 - How were the minimum and maximum volumes chosen? [I think he means how did you choose systole and diastole] Is it the min/max of the volume along the whole sequence or do you restrict the frame used?\r\n\r\nLet me know whether you find this helpful and I should plan on another one, subject to the PIs' availability.\r\n\r\n\r\n  [1]: https://www.kaggle.com/c/second-annual-data-science-bowl/forums/t/18198/q-a-with-principle-investigators-michael-hansen-ph-d-and-dr-andrew-arai",
    "104315": "Shannon, Thanks for putting this together.  And thanks to Dr. Hansen and Dr. Arai for their time.  I found the discussion informative and it gave me some new ideas to try with my model."
  },
  "source": "meta"
}