Zayif Etiketli Veriden Anlamsal Kavramlarin Oǧrenilmesi

Translated title of the contribution: Learning semantic concepts from weakly labeled data
  • Samet Hicsonmez
  • , Iman Rezazadeh
  • , Damla Unal
  • , Didem Yaniktepe
  • , Pinar Duygulu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

As in many areas with deep learning methods, great success is achieved in object and scene recognition, but for a good result a large number of labeled data is needed. Our aim in this study is to eliminate the need for tagged data collection that requires too much human labor, and to use the abundant images as data set, paired with the object name on the Internet and social media. However, this data is not as clean as manually labeled data, and it does not work well when used directly. In this study, Association with Model Evolution (AME) method is adapted to eliminate noisy data. Data that were automatically collected and cleaned with AME were then used as the experimental set for Convolutional Neural Networks (CNN). It is observed that the performance is increased by 4% with using the AME cleaned data.

Translated title of the contributionLearning semantic concepts from weakly labeled data
Original languageTurkish
Title of host publication27th Signal Processing and Communications Applications Conference, SIU 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728119045
DOIs
Publication statusPublished - Apr 2019
Event27th Signal Processing and Communications Applications Conference, SIU 2019 - Sivas, Turkey
Duration: 24 Apr 201926 Apr 2019

Publication series

Name27th Signal Processing and Communications Applications Conference, SIU 2019

Conference

Conference27th Signal Processing and Communications Applications Conference, SIU 2019
Country/TerritoryTurkey
CitySivas
Period24/04/1926/04/19

Fingerprint

Dive into the research topics of 'Learning semantic concepts from weakly labeled data'. Together they form a unique fingerprint.

Cite this