Journal article Open Access

Linear Maximum Margin Classifier for Learning from Uncertain Data

Christos Tzelepis; Vasileios Mezaris; Ioannis Patras


JSON Export

{
  "files": [
    {
      "links": {
        "self": "https://zenodo.org/api/files/3a50f4df-6764-4b38-91b7-f1d3836506c3/pami17_preprint.pdf"
      }, 
      "checksum": "md5:748edc5098c691cbbf0eef5fb9d5ccce", 
      "bucket": "3a50f4df-6764-4b38-91b7-f1d3836506c3", 
      "key": "pami17_preprint.pdf", 
      "type": "pdf", 
      "size": 2268913
    }
  ], 
  "owners": [
    22754
  ], 
  "doi": "10.1109/TPAMI.2017.2772235", 
  "stats": {
    "version_unique_downloads": 48.0, 
    "unique_views": 72.0, 
    "views": 82.0, 
    "version_views": 82.0, 
    "unique_downloads": 48.0, 
    "version_unique_views": 72.0, 
    "volume": 111176737.0, 
    "version_downloads": 49.0, 
    "downloads": 49.0, 
    "version_volume": 111176737.0
  }, 
  "links": {
    "doi": "https://doi.org/10.1109/TPAMI.2017.2772235", 
    "latest_html": "https://zenodo.org/record/1135049", 
    "bucket": "https://zenodo.org/api/files/3a50f4df-6764-4b38-91b7-f1d3836506c3", 
    "badge": "https://zenodo.org/badge/doi/10.1109/TPAMI.2017.2772235.svg", 
    "html": "https://zenodo.org/record/1135049", 
    "latest": "https://zenodo.org/api/records/1135049"
  }, 
  "created": "2018-01-04T11:28:58.448507+00:00", 
  "updated": "2020-01-20T17:21:52.556550+00:00", 
  "conceptrecid": "1135048", 
  "revision": 9, 
  "id": 1135049, 
  "metadata": {
    "access_right_category": "success", 
    "doi": "10.1109/TPAMI.2017.2772235", 
    "version": "v1", 
    "language": "eng", 
    "title": "Linear Maximum Margin Classifier for Learning from Uncertain Data", 
    "license": {
      "id": "CC-BY-4.0"
    }, 
    "journal": {
      "title": "IEEE Transactions on Pattern Analysis and Machine Intelligence"
    }, 
    "relations": {
      "version": [
        {
          "count": 1, 
          "index": 0, 
          "parent": {
            "pid_type": "recid", 
            "pid_value": "1135048"
          }, 
          "is_last": true, 
          "last_child": {
            "pid_type": "recid", 
            "pid_value": "1135049"
          }
        }
      ]
    }, 
    "communities": [
      {
        "id": "moving-h2020"
      }
    ], 
    "grants": [
      {
        "code": "693092", 
        "links": {
          "self": "https://zenodo.org/api/grants/10.13039/501100000780::693092"
        }, 
        "title": "Training towards a society of data-savvy information professionals to enable open leadership innovation", 
        "acronym": "MOVING", 
        "program": "H2020", 
        "funder": {
          "doi": "10.13039/501100000780", 
          "acronyms": [], 
          "name": "European Commission", 
          "links": {
            "self": "https://zenodo.org/api/funders/10.13039/501100000780"
          }
        }
      }
    ], 
    "keywords": [
      "Classification", 
      "Convex optimization", 
      "Gaussian anisotropic uncertainty", 
      "Large margin methods", 
      "Learning with uncertainty", 
      "Statistical learning theory"
    ], 
    "publication_date": "2017-12-29", 
    "creators": [
      {
        "affiliation": "School of Electronic Engineering and Computer Science, Queen Mary University of London", 
        "name": "Christos Tzelepis"
      }, 
      {
        "affiliation": "Informatics and Telematics Institute, Centre for Research and Technology Hellas", 
        "name": "Vasileios Mezaris"
      }, 
      {
        "affiliation": "EECS, Queen Mary University of London, London", 
        "name": "Ioannis Patras"
      }
    ], 
    "access_right": "open", 
    "resource_type": {
      "subtype": "article", 
      "type": "publication", 
      "title": "Journal article"
    }, 
    "description": "<p>In this paper, we propose a maximum margin classifier that deals with uncertainty in data input. More specifically, we reformulate the SVM framework such that each training example can be modeled by a multi-dimensional Gaussian distribution described by its mean vector and its covariance matrix -- the latter modeling the uncertainty. We address the classification problem and define a cost function that is the expected value of the classical SVM cost when data samples are drawn from the multi-dimensional Gaussian distributions that form the set of the training examples. Our formulation approximates the classical SVM formulation when the training examples are isotropic Gaussians with variance tending to zero. We arrive at a convex optimization problem which we solve efficiently in the primal form using a stochastic gradient descent approach. The resulting classifier, which we name SVM with Gaussian Sample Uncertainty (SVM-GSU), is tested on synthetic data and five publicly available and popular datasets; namely, the MNIST, WDBC, DEAP, TV News Channel Commercial Detection, and TRECVID MED datasets. Experimental results verify the effectiveness of the proposed method.</p>"
  }
}
82
49
views
downloads
Views 82
Downloads 49
Data volume 111.2 MB
Unique views 72
Unique downloads 48

Share

Cite as