Area: Deep Learning

Document Outline

• Abstract
• Abstract (Spanish)
• Index
• List of figures
• List of tables
• Introduction
  • Problem's description, relevance and justification
    • Zenodo and the consequences of spam
    • Zenodo's spam detection mechanism
    • Zenodo's spam taxonomy: learnt from experience
    • Conclusion
  • Motivation
  • Objectives
  • Methodology
  • Planning
    • Task groups description
    • Gantt chart
• State of the art
  • Classical methods
  • Feature extraction
    • Content based
    • User or behaviour based
    • Link based
  • Neural networks and deep learning approaches
    • CNN and RNN
  • Summary
• Design and implementation
  • Descriptive Analysis
    • Dataset description
  • Classical methods
    • Feature extraction
    • Random Forest based models
    • Conclusions and future work
  • Neural Networks
    • Neural Networks form the literature
    • CRNN on English-only content
    • Deeper Neural Networks
    • Testing the whole dataset
• Final conclusions and future work
  • Conclusion
  • Future work
• Deployment in production
• Bibliografía
• Deep Neural Networks configuration.