{
  "id": "2105.09270",
  "version": "v1",
  "published": "2021-05-19T17:30:28.000Z",
  "updated": "2021-05-19T17:30:28.000Z",
  "title": "Do We Really Need to Learn Representations from In-domain Data for Outlier Detection?",
  "authors": [
    "Zhisheng Xiao",
    "Qing Yan",
    "Yali Amit"
  ],
  "categories": [
    "cs.CV",
    "cs.LG"
  ],
  "abstract": "Unsupervised outlier detection, which predicts if a test sample is an outlier or not using only the information from unlabelled inlier data, is an important but challenging task. Recently, methods based on the two-stage framework achieve state-of-the-art performance on this task. The framework leverages self-supervised representation learning algorithms to train a feature extractor on inlier data, and applies a simple outlier detector in the feature space. In this paper, we explore the possibility of avoiding the high cost of training a distinct representation for each outlier detection task, and instead using a single pre-trained network as the universal feature extractor regardless of the source of in-domain data. In particular, we replace the task-specific feature extractor by one network pre-trained on ImageNet with a self-supervised loss. In experiments, we demonstrate competitive or better performance on a variety of outlier detection benchmarks compared with previous two-stage methods, suggesting that learning representations from in-domain data may be unnecessary for outlier detection.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2021-05-19T17:30:28.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "outlier detection",
      "in-domain data",
      "learn representations",
      "self-supervised representation learning algorithms",
      "feature extractor"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}