{
  "fetched_at": "2026-10-19T19:18:59Z",
  "source_url": "https://www.kaggle.com/competitions/bengali-ged/data",
  "pages": [
    {
      "name": "data-description",
      "content": "## FIles\nDataSetFold1.csv - training data\ntest.csv - test data\nSampleSubmission.csv - a sample submission file in the correct format\n## Columns\n* sentence - the erroneous sentence \n* gt - '$' signs are the markers for each error ",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "On the occasion of International language month, what better way to automate unsolved Bangla Grammatical Error Detection through Artificial Intelligence?\nEEE Day 2023 and Bengali.Ai are organizing \"Apurba presents Bhashabhrom\": Bangla Grammar error detection Datathon. The competition will be launched on a virtual platform followed by multiple hybrid and online workshops.\nEvent Timeline: 10 - 28 February\nParticipant requirement: 1-4 in one group. Students or young professionals (everyone must be current student/alumni of a Bangladeshi University) can form a team\nEvent platform: Kaggle\nOrganizer: EEE Day 2023, Bengali.Ai\nTechnical support: Banan Andolon\nSponsored by: Apurba Technologies\nRegister at shorturl.at/ezBLZ \nMore details [here](https://www.facebook.com/events/1283706012488022).\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "# **Team guidelines**\n- The maximum team size is 4\n- For overseas participants, a minimum of 1 person should be present in Bangladesh during the In-Person Evaluation session.\n- In parallel with the main prizepool, 'EmergingDataheads' is a separate prizepool for encouraging undergrad students.\n- To be considered as ‘EmergingDataheads’, the team should consist of undergrad students.\n- **Deadline for team merge is February 20, 2023**\n- No private code sharing outside teams\n- Internet usage is allowed for downloading python modules and models. No external grammar checker API are allowed.\n\n# **Evaluation**\nThe Public Leaderboard score is for benchmarking and validating participants' models and will also count toward the final evaluation score. A full evaluation will be based on the following factors (these are subject to change based on the organizers’ decision. Also, the weights of these factors will not be revealed to the participants):\n- Public score\n- Hidden Score\n- Reproducibility\n- Inference time\n- A report in IEEE Conference Paper Format\n- Presentation\n\nHidden Score: A separate test data will be preserved for the competition organizers that will be run through the models to get the hidden score\n\n# **Competition Data**\n- DATA ACCESS AND USE: Competition Use, Academic Use, and Non-Commercial Use.\n- If any external datasets are used those datasets have to be disclosed.\n\n# **General Rules**\n- The participants must only use open source code licensed under an Open Source Initiative-approved license (see www.opensource.org) that in no event limits commercial use of such code or model containing or depending on such code.\n- Winner license: The winners will grant the competition hosts and sponsors the following license: \nYou hereby license and will license your winning submission and the source code used to generate the submission under an Open Source Initiative-approved license (see www.opensource.org) that in no event limits commercial use of such code or model containing or depending on such code. For generally commercially available software that you used to generate the submission that is not owned by you but that can be procured by the Competition Sponsor without undue expense, you do not need to grant the license in the preceding sentence for that software. You must also make your solution public.\n- The Competition hosts and sponsors and its affiliates may use your name and likeness for advertising and promotional purposes without additional compensation unless prohibited by law.\n- You warrant that your submission is your own original work and, as such, you are the sole and exclusive owner and rights holder of the submission, and you have the right to make the submission and grant all required licenses. To the maximum extent permitted by law, you indemnify and agree to keep indemnified Competition Entities at all times from and against any liability, claims, demands, losses, damages, costs, and expenses resulting from any of your acts, defaults or omissions and/or a breach of any warranty set forth herein.\n- **The Competition Organizing Committee reserves the right to cancel or postpone the event for any reason.**\n- **The organizers serve the right to disqualify any team.**\n- **Winners will be selected at the sole discretion of the Competition Organizing Committee. All participants must conform to the terms and conditions. The Competition Organizing Committee’s decisions in this regard will be final, and no queries or appeals thereof will be entertained.**\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Bengali Grammatical Error Detection",
      "content": "The task is to detect substrings that contain grammatical or spelling error.\n\nIn the sentence \"আমী ভাত খাই।\", the first word is erroneous. So we flag the word only and return \"$আমী$ ভাত খাই।\" The participants have to detect all the substrings and provide the predicted string with '$' marks.\n\nThe major error types are:\n1. Single word error অনেক $বাতাষ$ হচ্ছে।\n2. Multi word error $বাংলা ধেষ$ আমাদের জন্মভূমি।\n3. Wrong Punctuation আপনি কি আমাকে দেখেছেন$।$\n4. Punctuation Omission আপনি কি রিক অ্যান্ড মর্টি দেখেছেন$$ (there will be a punctuation that is not present. So we mark the position of the error with two $ signs)\n5. Merge Error $ভাললাগছে$ ক্লাস করতে।\n6. Form/Inflection Error\n7. Unwanted space error তুমি ভাত খাও$ $।\n8. Hybrid\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Timeline",
      "content": "- Team Merger deadline: **February 25, 2023.**\n- Release of the remaining train data: **February 22, 2023.**\n- Phase 1 Submission Deadline: **February 28, 2023.**\n- Notebook Submission Deadline: **March 3, 2023.**\n- Phase 2 (Offline Presentation): **First Week of March (tentative).**",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "For the first round, the Kaggle leaderboard will be based on [Levenshtein Distance](https://en.wikipedia.org/wiki/Levenshtein_distance) of the predicted string and ground truth. After the end of the first round (28th February), all the submissions will be evaluated using a separate held-out test set. The participants will have to submit their Training and inference Notebook as well as a detailed paper on their approach after the first phase ends.\n\nThe final Leaderboard will be decided based on the first-phase score on the partial test set (20%), the second-phase score on the held-out test set (50%), Runtime (10%), and the Paper + Final Presentation (20%).\n\n##### Demonstration:\n\n\nSuppose a given sentence is \"আপণা কে দেখলাম সেদিন রাস্তায়\"\nThere is an error at the beginning of the sentence : \"আপণা কে\" > \"আপনাকে\"\n\nAlso, there should also be punctuation at the end of the sentence.\n\nSo, the ground truth should be >\"$আপণা কে$ দেখলাম সেদিন রাস্তায়$$\"\n\nNow, if the prediction is :  \"আপণা কে দেখলাম $সেদিন$ রাস্তায়\",\n the Levenshtein Distance between the ground truth and the prediction is >lev(\"$আপণা কে$ দেখলাম সেদিন রাস্তায়$$\", \"আপণা কে দেখলাম $সেদিন$ রাস্তায়\").",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}