[
  {
    "Percentage error": 5.5,
    "code_links": [],
    "date": "2017-03-06",
    "date2": 20170306,
    "model": "ResNet + BiLSTMs acoustic model",
    "paper": {
      "title": "English Conversational Telephone Speech Recognition by Humans and Machines",
      "url": "https://cknow.io/lib/d65a228a6d6d00a9"
    },
    "paper_data_uoa": "d65a228a6d6d00a9"
  },
  {
    "Percentage error": 5.8,
    "code_links": [],
    "date": "2016-10-17",
    "date2": 20161017,
    "model": "Microsoft 2016b",
    "paper": {
      "title": "Achieving Human Parity in Conversational Speech Recognition",
      "url": "https://cknow.io/lib/7f54ea194dbda268"
    },
    "paper_data_uoa": "7f54ea194dbda268"
  },
  {
    "Percentage error": 6.2,
    "code_links": [],
    "date": "2016-09-12",
    "date2": 20160912,
    "model": "Microsoft 2016",
    "paper": {
      "title": "The Microsoft 2016 Conversational Speech Recognition System",
      "url": "https://cknow.io/lib/1ba47009467bdc18"
    },
    "paper_data_uoa": "1ba47009467bdc18"
  },
  {
    "Percentage error": 6.3,
    "code_links": [],
    "date": "2016-09-12",
    "date2": 20160912,
    "model": "VGG/Resnet/LACE/BiLSTM acoustic model trained on SWB+Fisher+CH, N-gram + RNNLM language model trained on Switchboard+Fisher+Gigaword+Broadcast",
    "paper": {
      "title": "The Microsoft 2016 Conversational Speech Recognition System",
      "url": "https://cknow.io/lib/1ba47009467bdc18"
    },
    "paper_data_uoa": "1ba47009467bdc18"
  },
  {
    "Percentage error": 6.6,
    "code_links": [],
    "date": "2016-04-27",
    "date2": 20160427,
    "model": "RNN + VGG + LSTM acoustic model trained on SWB+Fisher+CH, N-gram + \"model M\" + NNLM language model",
    "paper": {
      "title": "The IBM 2016 English Conversational Telephone Speech Recognition System",
      "url": "https://cknow.io/lib/68870341a9802e1a"
    },
    "paper_data_uoa": "68870341a9802e1a"
  },
  {
    "Percentage error": 6.6,
    "code_links": [],
    "date": "2016-10-17",
    "date2": 20161017,
    "model": "CNN-LSTM",
    "paper": {
      "title": "Achieving Human Parity in Conversational Speech Recognition",
      "url": "https://cknow.io/lib/7f54ea194dbda268"
    },
    "paper_data_uoa": "7f54ea194dbda268"
  },
  {
    "Percentage error": 6.8,
    "code_links": [
      {
        "title": "mozilla/DeepSpeech",
        "url": "https://github.com/mozilla/DeepSpeech"
      },
      {
        "title": "shelling203/SpecAugment",
        "url": "https://github.com/shelling203/SpecAugment"
      },
      {
        "title": "DemisEom/SpecAugment",
        "url": "https://github.com/DemisEom/SpecAugment"
      },
      {
        "title": "lRomul/argus-freesound",
        "url": "https://github.com/lRomul/argus-freesound"
      },
      {
        "title": "ebouteillon/freesound-audio-tagging-2019",
        "url": "https://github.com/ebouteillon/freesound-audio-tagging-2019"
      },
      {
        "title": "sh951011/Korean-Speech-Recognition",
        "url": "https://github.com/sh951011/Korean-Speech-Recognition/blob/master/package/feature.py"
      },
      {
        "title": "sh951011/Korean-Speech-Recognition",
        "url": "https://github.com/sh951011/Korean-Speech-Recognition"
      },
      {
        "title": "KimJeongSun/SpecAugment_numpy_scipy",
        "url": "https://github.com/KimJeongSun/SpecAugment_numpy_scipy"
      },
      {
        "title": "hgstudent/las",
        "url": "https://github.com/hgstudent/las"
      },
      {
        "title": "viig99/mixmatch-freesound",
        "url": "https://github.com/viig99/mixmatch-freesound"
      },
      {
        "title": "knlee-voice/PaperNotes",
        "url": "https://github.com/knlee-voice/PaperNotes"
      }
    ],
    "date": "2019-04-18",
    "date2": 20190418,
    "model": "LAS + SpecAugment (SM)",
    "paper": {
      "title": "SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition",
      "url": "https://cknow.io/lib/15408ea14d4c82f9"
    },
    "paper_data_uoa": "15408ea14d4c82f9"
  },
  {
    "Percentage error": 6.9,
    "code_links": [],
    "date": "2016-04-27",
    "date2": 20160427,
    "model": "IBM 2016",
    "paper": {
      "title": "The IBM 2016 English Conversational Telephone Speech Recognition System",
      "url": "https://cknow.io/lib/68870341a9802e1a"
    },
    "paper_data_uoa": "68870341a9802e1a"
  },
  {
    "Percentage error": 6.9,
    "code_links": [],
    "date": "2016-09-12",
    "date2": 20160912,
    "model": "RNNLM",
    "paper": {
      "title": "The Microsoft 2016 Conversational Speech Recognition System",
      "url": "https://cknow.io/lib/1ba47009467bdc18"
    },
    "paper_data_uoa": "1ba47009467bdc18"
  },
  {
    "Percentage error": 8,
    "code_links": [],
    "date": "2015-05-21",
    "date2": 20150521,
    "model": "IBM 2015",
    "paper": {
      "title": "The IBM 2015 English Conversational Telephone Speech Recognition System",
      "url": "https://cknow.io/lib/4d8bf3030d142d17"
    },
    "paper_data_uoa": "4d8bf3030d142d17"
  },
  {
    "Percentage error": 12.2,
    "code_links": [],
    "date": "2015-09-29",
    "date2": 20150929,
    "model": "Deep CNN (10 conv, 4 FC layers), multi-scale feature maps",
    "paper": {
      "title": "Very Deep Multilingual Convolutional Neural Networks for LVCSR",
      "url": "https://cknow.io/lib/bfdb5a624b4b9299"
    },
    "paper_data_uoa": "bfdb5a624b4b9299"
  },
  {
    "Percentage error": 12.6,
    "code_links": [
      {
        "title": "mozilla/DeepSpeech",
        "url": "https://github.com/mozilla/DeepSpeech"
      },
      {
        "title": "robmsmt/KerasDeepSpeech",
        "url": "https://github.com/robmsmt/KerasDeepSpeech"
      },
      {
        "title": "myrtleSoftware/deepspeech",
        "url": "https://github.com/myrtleSoftware/deepspeech"
      },
      {
        "title": "IBM/MAX-Speech-to-Text-Converter",
        "url": "https://github.com/IBM/MAX-Speech-to-Text-Converter"
      },
      {
        "title": "mangushev/deep_speech",
        "url": "https://github.com/mangushev/deep_speech"
      },
      {
        "title": "RashadGarayev/TRSpeech-to-text",
        "url": "https://github.com/RashadGarayev/TRSpeech-to-text"
      },
      {
        "title": "WalterJohnson0/DeepSpeech-KerasRebuild",
        "url": "https://github.com/WalterJohnson0/DeepSpeech-KerasRebuild"
      },
      {
        "title": "Loghijiaha/DeepSpeech-Indo",
        "url": "https://github.com/Loghijiaha/DeepSpeech-Indo"
      }
    ],
    "date": "2014-12-17",
    "date2": 20141217,
    "model": "Deep Speech + FSH",
    "paper": {
      "title": "Deep Speech: Scaling up end-to-end speech recognition",
      "url": "https://cknow.io/lib/23ae74412eb5aaa6"
    },
    "paper_data_uoa": "23ae74412eb5aaa6"
  },
  {
    "Percentage error": 12.6,
    "code_links": [
      {
        "title": "mozilla/DeepSpeech",
        "url": "https://github.com/mozilla/DeepSpeech"
      },
      {
        "title": "robmsmt/KerasDeepSpeech",
        "url": "https://github.com/robmsmt/KerasDeepSpeech"
      },
      {
        "title": "myrtleSoftware/deepspeech",
        "url": "https://github.com/myrtleSoftware/deepspeech"
      },
      {
        "title": "IBM/MAX-Speech-to-Text-Converter",
        "url": "https://github.com/IBM/MAX-Speech-to-Text-Converter"
      },
      {
        "title": "mangushev/deep_speech",
        "url": "https://github.com/mangushev/deep_speech"
      },
      {
        "title": "RashadGarayev/TRSpeech-to-text",
        "url": "https://github.com/RashadGarayev/TRSpeech-to-text"
      },
      {
        "title": "WalterJohnson0/DeepSpeech-KerasRebuild",
        "url": "https://github.com/WalterJohnson0/DeepSpeech-KerasRebuild"
      },
      {
        "title": "Loghijiaha/DeepSpeech-Indo",
        "url": "https://github.com/Loghijiaha/DeepSpeech-Indo"
      }
    ],
    "date": "2014-12-17",
    "date2": 20141217,
    "model": "CNN + Bi-RNN + CTC (speech to letters), 25.9% WER if trainedonlyon SWB",
    "paper": {
      "title": "Deep Speech: Scaling up end-to-end speech recognition",
      "url": "https://cknow.io/lib/23ae74412eb5aaa6"
    },
    "paper_data_uoa": "23ae74412eb5aaa6"
  },
  {
    "Percentage error": 15,
    "code_links": [],
    "date": "2014-06-30",
    "date2": 20140630,
    "model": "DNN + Dropout",
    "paper": {
      "title": "Building DNN Acoustic Models for Large Vocabulary Speech Recognition",
      "url": "https://cknow.io/lib/1be14a5be24c6385"
    },
    "paper_data_uoa": "1be14a5be24c6385"
  },
  {
    "Percentage error": 16,
    "code_links": [],
    "date": "2014-06-30",
    "date2": 20140630,
    "model": "DNN",
    "paper": {
      "title": "Building DNN Acoustic Models for Large Vocabulary Speech Recognition",
      "url": "https://cknow.io/lib/1be14a5be24c6385"
    },
    "paper_data_uoa": "1be14a5be24c6385"
  },
  {
    "Percentage error": 20,
    "code_links": [
      {
        "title": "mozilla/DeepSpeech",
        "url": "https://github.com/mozilla/DeepSpeech"
      },
      {
        "title": "robmsmt/KerasDeepSpeech",
        "url": "https://github.com/robmsmt/KerasDeepSpeech"
      },
      {
        "title": "myrtleSoftware/deepspeech",
        "url": "https://github.com/myrtleSoftware/deepspeech"
      },
      {
        "title": "IBM/MAX-Speech-to-Text-Converter",
        "url": "https://github.com/IBM/MAX-Speech-to-Text-Converter"
      },
      {
        "title": "mangushev/deep_speech",
        "url": "https://github.com/mangushev/deep_speech"
      },
      {
        "title": "RashadGarayev/TRSpeech-to-text",
        "url": "https://github.com/RashadGarayev/TRSpeech-to-text"
      },
      {
        "title": "WalterJohnson0/DeepSpeech-KerasRebuild",
        "url": "https://github.com/WalterJohnson0/DeepSpeech-KerasRebuild"
      },
      {
        "title": "Loghijiaha/DeepSpeech-Indo",
        "url": "https://github.com/Loghijiaha/DeepSpeech-Indo"
      }
    ],
    "date": "2014-12-17",
    "date2": 20141217,
    "model": "Deep Speech",
    "paper": {
      "title": "Deep Speech: Scaling up end-to-end speech recognition",
      "url": "https://cknow.io/lib/23ae74412eb5aaa6"
    },
    "paper_data_uoa": "23ae74412eb5aaa6"
  }
]