{
  "id": 212830,
  "title": "Test-time augmentation (or TTA)",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/212830",
  "author_name": "",
  "post_date": "2021-01-20T11:10:03.860362300Z",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<h1>Test-time augmentation (or TTA)</h1>\n<p>Test-time augmentation (TTA)—the aggregation of predictions across transformed versions of a test input—is a common practice in image classification.</p>\n<h2>Blogs:</h2>\n<h3>How to Use Test-Time Augmentation to Make Better Predictions:</h3>\n<p><a href=\"https://machinelearningmastery.com/how-to-use-test-time-augmentation-to-improve-model-performance-for-image-classification/\" target=\"_blank\">https://machinelearningmastery.com/how-to-use-test-time-augmentation-to-improve-model-performance-for-image-classification/</a></p>\n<h3>Test-Time Augmentation For Tabular Data With Scikit-Learn:</h3>\n<p><a href=\"https://machinelearningmastery.com/test-time-augmentation-with-scikit-learn/\" target=\"_blank\">https://machinelearningmastery.com/test-time-augmentation-with-scikit-learn/</a></p>\n<h2>Papers:</h2>\n<h3>When and Why Test-Time Augmentation Works</h3>\n<p><a href=\"https://arxiv.org/pdf/2011.11156.pdf\" target=\"_blank\">https://arxiv.org/pdf/2011.11156.pdf</a></p>\n<h3>Test-time augmentation with uncertainty estimation for deep learning-based medical image segmentation</h3>\n<p><a href=\"https://openreview.net/pdf?id=Byxv9aioz\" target=\"_blank\">https://openreview.net/pdf?id=Byxv9aioz</a></p>\n<h3>Learning Loss for Test-Time Augmentation</h3>\n<p><a href=\"https://arxiv.org/pdf/2010.11422.pdf\" target=\"_blank\">https://arxiv.org/pdf/2010.11422.pdf</a></p>\n<h3>Greedy Policy Search: A Simple Baseline for Learnable Test-Time Augmentation</h3>\n<p><a href=\"https://arxiv.org/pdf/2002.09103.pdf\" target=\"_blank\">https://arxiv.org/pdf/2002.09103.pdf</a></p>",
  "messages": [
    {
      "id": "1161147",
      "postDate": "01/20/2021 11:10:03",
      "content": "<h1>Test-time augmentation (or TTA)</h1>\n<p>Test-time augmentation (TTA)—the aggregation of predictions across transformed versions of a test input—is a common practice in image classification.</p>\n<h2>Blogs:</h2>\n<h3>How to Use Test-Time Augmentation to Make Better Predictions:</h3>\n<p><a href=\"https://machinelearningmastery.com/how-to-use-test-time-augmentation-to-improve-model-performance-for-image-classification/\" target=\"_blank\">https://machinelearningmastery.com/how-to-use-test-time-augmentation-to-improve-model-performance-for-image-classification/</a></p>\n<h3>Test-Time Augmentation For Tabular Data With Scikit-Learn:</h3>\n<p><a href=\"https://machinelearningmastery.com/test-time-augmentation-with-scikit-learn/\" target=\"_blank\">https://machinelearningmastery.com/test-time-augmentation-with-scikit-learn/</a></p>\n<h2>Papers:</h2>\n<h3>When and Why Test-Time Augmentation Works</h3>\n<p><a href=\"https://arxiv.org/pdf/2011.11156.pdf\" target=\"_blank\">https://arxiv.org/pdf/2011.11156.pdf</a></p>\n<h3>Test-time augmentation with uncertainty estimation for deep learning-based medical image segmentation</h3>\n<p><a href=\"https://openreview.net/pdf?id=Byxv9aioz\" target=\"_blank\">https://openreview.net/pdf?id=Byxv9aioz</a></p>\n<h3>Learning Loss for Test-Time Augmentation</h3>\n<p><a href=\"https://arxiv.org/pdf/2010.11422.pdf\" target=\"_blank\">https://arxiv.org/pdf/2010.11422.pdf</a></p>\n<h3>Greedy Policy Search: A Simple Baseline for Learnable Test-Time Augmentation</h3>\n<p><a href=\"https://arxiv.org/pdf/2002.09103.pdf\" target=\"_blank\">https://arxiv.org/pdf/2002.09103.pdf</a></p>",
      "rawMarkdown": "# Test-time augmentation (or TTA)\n\nTest-time augmentation (TTA)—the aggregation of predictions across transformed versions of a test input—is a common practice in image classification.\n\n## Blogs:\n\n### How to Use Test-Time Augmentation to Make Better Predictions:\nhttps://machinelearningmastery.com/how-to-use-test-time-augmentation-to-improve-model-performance-for-image-classification/\n\n### Test-Time Augmentation For Tabular Data With Scikit-Learn:\nhttps://machinelearningmastery.com/test-time-augmentation-with-scikit-learn/\n\n## Papers:\n\n### When and Why Test-Time Augmentation Works\nhttps://arxiv.org/pdf/2011.11156.pdf\n\n### Test-time augmentation with uncertainty estimation for deep learning-based medical image segmentation\nhttps://openreview.net/pdf?id=Byxv9aioz\n\n### Learning Loss for Test-Time Augmentation\nhttps://arxiv.org/pdf/2010.11422.pdf\n\n### Greedy Policy Search: A Simple Baseline for Learnable Test-Time Augmentation\nhttps://arxiv.org/pdf/2002.09103.pdf",
      "votes": null
    },
    {
      "id": "1161148",
      "postDate": "01/20/2021 11:11:14",
      "content": "<p>Thanks for sharing!</p>",
      "rawMarkdown": "Thanks for sharing!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1161148,
      "author_name": "sajidhussain3",
      "author_url": "",
      "post_date": "01/20/2021 11:11:14",
      "content": "<p>Thanks for sharing!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1161147": "# Test-time augmentation (or TTA)\n\nTest-time augmentation (TTA)—the aggregation of predictions across transformed versions of a test input—is a common practice in image classification.\n\n## Blogs:\n\n### How to Use Test-Time Augmentation to Make Better Predictions:\nhttps://machinelearningmastery.com/how-to-use-test-time-augmentation-to-improve-model-performance-for-image-classification/\n\n### Test-Time Augmentation For Tabular Data With Scikit-Learn:\nhttps://machinelearningmastery.com/test-time-augmentation-with-scikit-learn/\n\n## Papers:\n\n### When and Why Test-Time Augmentation Works\nhttps://arxiv.org/pdf/2011.11156.pdf\n\n### Test-time augmentation with uncertainty estimation for deep learning-based medical image segmentation\nhttps://openreview.net/pdf?id=Byxv9aioz\n\n### Learning Loss for Test-Time Augmentation\nhttps://arxiv.org/pdf/2010.11422.pdf\n\n### Greedy Policy Search: A Simple Baseline for Learnable Test-Time Augmentation\nhttps://arxiv.org/pdf/2002.09103.pdf",
    "1161148": "Thanks for sharing!"
  },
  "source": "meta"
}