MLchartDataset catalogue

Double descent

Term · Oil and gas · MLC-T-OAG-001146

A phenomenon observed in machine learning where model performance initially improves with increasing model complexity, then degrades, and surprisingly improves again as complexity further increases, often beyond the point of perfect fit on training data. This challenges traditional understanding of the bias-variance trade-off. It is particularly relevant in overparameterized deep learning models.

Table 1. Record
IdentifierMLC-T-OAG-001146
FieldOil and gas
SubjectDigital
Record as JSON 
{
  "id": "MLC-T-OAG-001146",
  "term": "Double descent",
  "field": "Oil and gas",
  "definition": "A phenomenon observed in machine learning where model performance initially improves with increasing model complexity, then degrades, and surprisingly improves again as complexity further increases, often beyond the point of perfect fit on training data. This challenges traditional understanding of the bias-variance trade-off. It is particularly relevant in overparameterized deep learning models.",
  "subject": "Digital",
  "url": "https://mlchart.com/terminology/oil-and-gas/double-descent/"
}

Record 1,228 of 4,469 in Oil and gas terminology (MLC-0103). Request the full dataset.