{"id":2178,"title":"Singing Voice Deepfake Detection (CtrSVDD Track result submission)","published":true,"created_by":"yzyouzhang","owner_display_name":"yzyouzhang","created_when":"2024-03-03T03:47:07.677855Z","logo":"https://miniodis-rproxy.lisn.upsaclay.fr/coda-v2-prod-public/logos/2024-03-03-1709437627/9038d0b0b7d4/svdd.png?AWSAccessKeyId=codabench-prod&Signature=%2FF164DShM9AjrzkvujRSuMq0Q%2Fg%3D&Expires=1782600283","logo_icon":"https://miniodis-rproxy.lisn.upsaclay.fr/coda-v2-prod-public/logos/2024-03-03-1709437627/9038d0b0b7d4/svdd_icon.png?AWSAccessKeyId=codabench-prod&Signature=t6bEyCA3%2FfH%2B3KIRT%2F4SBcqIVsw%3D&Expires=1782600283","terms":"# Terms and Condition\n\nPlease refer to [our challenge website](https://challenge.singfake.org) for detailed information.\n\n\nThe CtrSVDD datasets we released are under under CC-BY-NC-ND 4.0 license, aligned with the sourcing corpora.\n\nTraining/Development set: [https://zenodo.org/records/10467648](https://zenodo.org/records/10467648).\n\nTest set: [https://zenodo.org/records/10742049](https://zenodo.org/records/10742049).","pages":[{"id":20971,"title":"Participation","content":"# How to Participate\n\n## The challenge has ended, and we have released our test set keys and labels [here](https://zenodo.org/records/12703261)!\n\nFor detailed information on the evaluation plan, please refer to our [SVDD challenge website](https://challenge.singfake.org).\n\nParticipants are required to submit their predictions in a text file named according to the format `{team name}_{trial number}.txt`. This file should contain your model's predictions on the test dataset available through [this Zenodo link](https://zenodo.org/records/10742049). The expected format for each line in the text file is as follows:\n\n```text\nCtrSVDD_0125_E_0000001 2.8742611408233643\nCtrSVDD_0114_E_0000002 2.9362313747406006\n\n...\n```\n\nEach line should list the filename (omitting the \".flac\" extension) of a singing voice utterance, followed by a space, and then the score assigned by your prediction model to that utterance. A higher score indicates a higher confidence level that the utterance is bona fide. There is no specific range for the scores; any floating-point number is acceptable. The primary metric for evaluation will be the Equal Error Rate (EER).\n\nWe have prepared two baseline submissions (under *yzyouzhang* and *yongyizang*). Please refer to our [evaluation plan](https://challenge.singfake.org/SVDD_Challenge_2024_Eval_Plan_v0.3.pdf) and [baseline system code](https://github.com/SVDDChallenge/CtrSVDD2024_Baseline) for more details.\n\nAfter preparing your text file, compress it into a ZIP file with an identical name to the text file (`{team name}_{trial number}.zip`).\n\n<!-- Each team can ONLY submit the predicted score files THREE times for the CtrSVDD challenge ranking. More scores submissions can be used for comparative studies in participants' research paper writing and the organizers will open a new CodaBench and approve your registration of that CodaBench after you have used up your THREE chances. Note that only the submissions in this current CodaBench will count towards the challenge results/ranking. -->\n\nParticipants are limited to ONLY THREE submissions of predicted scores for each utterance within the scope of the CtrSVDD challenge for official ranking purposes. Additional submissions may be utilized for comparative analyses in participants' research papers. Upon exhausting your THREE submission opportunities, the organizers will facilitate access to a new CodaBench for further submissions, which will only be considered for research purposes and not for official challenge ranking. It's important to note that CodaBench's daily submission limit is separate; our three-submission limit refers to the total allowable submissions for the challenge.\n\nFor information regarding competition timelines, please visit the \"Timeline\" section of our website.","index":0}],"phases":[{"id":3698,"index":0,"start":"2024-03-02T00:00:00Z","end":"2024-06-19T00:00:00Z","name":"CtrSVDD Track","description":"SVDD Challenge (CtrSVDD track): submit your scores.","status":"Previous","execution_time_limit":600,"tasks":[{"task":"91cd34cb-d8e4-4e5d-8ce9-f61b902e3774","order_index":0,"phase":3698,"id":5494,"value":"91cd34cb-d8e4-4e5d-8ce9-f61b902e3774","key":"91cd34cb-d8e4-4e5d-8ce9-f61b902e3774","created_when":"2024-03-06T20:42:49.084529Z","name":"CtrSVDD Track EER","solutions":[],"public_datasets":[{"id":52429,"created_by":"yongyizang","owner_display_name":"yongyizang","created_when":"2024-03-06T20:06:14.651172Z","name":"reference data 2","type":"reference_data","description":"","is_public":false,"key":"e5f1a49f-0d28-41cb-8daf-c9ec871a97ee","value":"e5f1a49f-0d28-41cb-8daf-c9ec871a97ee","was_created_by_competition":false,"in_use":[{"pk":2177,"title":"Singing Voice Deepfake Detection (CtrSVDD Track result submission)"},{"pk":2178,"title":"Singing Voice Deepfake Detection (CtrSVDD Track result submission)"}],"competition":null,"file_name":"reference_data.zip","file_size":475473.92},{"id":52430,"created_by":"yongyizang","owner_display_name":"yongyizang","created_when":"2024-03-06T20:18:31.968153Z","name":"new scoring","type":"scoring_program","description":"","is_public":false,"key":"0fb7caf6-d34c-41ca-b941-ee9578b71573","value":"0fb7caf6-d34c-41ca-b941-ee9578b71573","was_created_by_competition":false,"in_use":[{"pk":2177,"title":"Singing Voice Deepfake Detection (CtrSVDD Track result submission)"},{"pk":2178,"title":"Singing Voice Deepfake Detection (CtrSVDD Track result submission)"}],"competition":null,"file_name":"scoring_program.zip","file_size":478289.92}]}],"has_max_submissions":true,"max_submissions_per_day":3,"max_submissions_per_person":3,"auto_migrate_to_this_phase":false,"hide_output":false,"hide_prediction_output":false,"hide_score_output":false,"public_data":null,"starting_kit":null,"is_final_phase":true,"used_submissions_per_day":0,"used_submissions_per_person":0}],"leaderboards":[{"id":2258,"primary_index":0,"title":"Results","key":"main","columns":[{"id":6339,"computation":null,"computation_indexes":null,"title":"Prediction EER","key":"pooled_attack_discard_A14_eer","sorting":"asc","index":0,"hidden":false,"precision":4}],"hidden":false,"submission_rule":"Add"}],"collaborators":[{"id":3570,"name":null,"username":"yongyizang"}],"registration_auto_approve":true,"description":"Controlled Singing Voice Deepfake Detection Task","participants_count":83,"submissions_count":130,"queue":null,"enable_detailed_results":false,"show_detailed_results_in_submission_panel":true,"show_detailed_results_in_leaderboard":true,"auto_run_submissions":true,"can_participants_make_submissions_public":true,"make_programs_available":true,"make_input_data_available":false,"docker_image":"codalab/codalab-legacy:py37","allow_robot_submissions":false,"competition_type":"competition","fact_sheet":null,"forum":2097,"reward":"","contact_email":"svddchallenge@gmail.com","report":"https://challenge.singfake.org/","forum_enabled":true}