[
  {
    "file": "paper_46.txt",
    "start": 410,
    "end": 415,
    "label": "Unsupported claim",
    "user": "Ed",
    "text": "LSTM "
  },
  {
    "file": "paper_46.txt",
    "start": 201,
    "end": 231,
    "label": "Unsupported claim",
    "user": "Ed",
    "text": "conditional random field (CRF)"
  },
  {
    "file": "paper_46.txt",
    "start": 709,
    "end": 824,
    "label": "Unsupported claim",
    "user": "Ed",
    "text": "Nested NER allows a token to belong to multiple entities, which conflicts with the plain sequence tagging framework"
  },
  {
    "file": "paper_46.txt",
    "start": 826,
    "end": 1280,
    "label": "Coherence",
    "user": "Ed",
    "text": "Ju et al. (2018) proposed to use stacked LSTM-CRFs to predict from inner to outer entities. Straková et al. (2019) concatenated the BILOU tags for each token inside the nested entities, which allows the LSTM-CRF to work as for flat entities. Li et al. (2020b) reformulated nested NER as a machine reading comprehension task. Shen et al. (2021) proposed to recognize nested entities by the two-stage object detection method widely used in computer vision."
  },
  {
    "file": "paper_46.txt",
    "start": 2065,
    "end": 2736,
    "label": "Lacks synthesis",
    "user": "Ed",
    "text": "Label Smoothing Szegedy et al. (2016) proposed the label smoothing as a regularization technique to improve the accuracy of the Inception networks on ImageNet. By explicitly assigning a small probability to non-ground-truth labels, label smoothing can prevent the models from becoming too confident about the predictions, and thus improve generalization. It turned out to be a useful alternative to the standard cross entropy loss, and has been widely adopted to fight against the over-confidence (Zoph et al., 2018;Chorowski and Jaitly, 2017;Vaswani et al., 2017), improve the model calibration (Müller et al., 2019), and denoise incorrect labels (Lukasik et al., 2020)."
  },
  {
    "file": "paper_46.txt",
    "start": 2844,
    "end": 2969,
    "label": "Unsupported claim",
    "user": "Ed",
    "text": "This is driven by the observation that entity boundaries are more ambiguous and inconsistent to annotate in NER engineering. "
  }
]