PRobability Mass Estimation in Tensors with Hidden Elements Using Structure

Project Description

The aim of project is to estimate the probability mass function for tensors with hidden elements using structure. This involves the development of new tools for learning the statistical behavior of discrete random vectors from their partial realizations. The joint probability mass function (PMF) of the discrete elements of the random vector takes the form of a multidimensional tensor.

• Objectives:
  • Designing new estimation algorithms
  • Designing new model order estimation methods
  • Derivation of theoretical estimation bounds
  • Exploring alternative statistical models
  • Experimenting with real data
• Application areas: There are variety of applications for statistical inference tasks
  • Data fusion (e.g. in biomedical applications)
    • computer-aided diagnosis
  • Recommender systems for movies, restaurants, hotels, etc.
  • Automated recruitment in recruitment processes
  • University admission systems
• Project duration: 3 years (2022-2025)
• Principal Investigators
  • Univ.-Prof. Dr.-Ing. Martin Haardt
  • Prof. A. Yeredor (Tel Aviv University, Israel)