{
  "fetched_at": "2026-10-19T06:28:48Z",
  "source_url": "https://www.kaggle.com/competitions/first-arrivals-ml-in-physics/data",
  "pages": [
    {
      "name": "data-description",
      "content": "There are 24 simulations in the train folder. each simulation folder contains 127 csv files, standing for 127 receivers. Each receiver records fluctuations in X, Y and Z components over a 30000 timesteps period. There can be from 0 (minimum) to 5 (maximum) earthquakes in each simulation. Your task is to determine when a receiver first picks up the energy from an earthquake. \n\nFor each simulation, there is a \"simulation_[X]_output.csv\". In this file, there is a 5 by 127 matrix, where each row corresponds for an \"event\" (an earthquake), and each column represents a receiver respectively. Each value is the exact timestep when the receiver recognises the first arrival of an earthquake. If there is less than 5 earthquakes in a simulation, you must fill everything else with -1. \n\nDue to Kaggle restrictions, the submission file cannot be in the matrix form but in a flattened format. Please refer to the sampleSubmission.csv to understand the layout. The file contains 127 x 5 =  635 entries, where each entry is in the format R[receiver #]-E[event #], [predicted timestep].",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "The goal of this challenge is to determine the First Arrivals of earthquakes. Before formulating our specific problem, let us consider a short example of a similar problem which may help to understand the general idea.\n\nWe have a smartphone and we want to ask an audio assistant (e.g. \"Ok Google!\") to find some information on\nthe internet.\n\nThe audio assistant has to continuously record our voice from the microphone and when we speak the keywords\n\"Ok Google!\", it activates speech recognition. To catch the keywords, the program inside of your smartphone\nhas to identify your voice among background noises and understand when your speech starts and ends. This\nis called voice activity detection. \n\nLet us come back to our problem, where \"voice\" is earthquakes and \"microphone\" is geophones recording quakes (receiver for short). Instead of detecting full voice activity, we want to identify only when it starts. In seismology, this procedure is called first-arrival picking. In other words, we want to understand when an earthquake starts in time.\n\nWe will use Mean Absolute Error (MAE) to evaluate the performance of the submitted prediction results and rank the best-performed models.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "Rules and detailed information can be found here: https://docs.google.com/presentation/d/1-xqRvz8G_ZVHMUgI-yMbz3aFVGFpq5rvQAYKkXruIh0/edit?usp=sharing\n\nPlease make sure you and all your teammates (if any) are registered through the google form: https://forms.gle/aVaB4vQHaANyzA9U7",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "Similarly to the first challenge, this challenge will also use MAE to determine the scoreboard. ",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}