STRONGLY CORRELATED INTERACTING PARTICLE SYSTEMS AND THEIR APPLICATIONS

Strongly correlated interacting particle systems are probabilistic models that describe collections of particles whose behaviors are mutually dependent, typically due to repulsive or attractive forces. These models have long been central to statistical physics, where they are used to describe gases, plasmas, or spin systems. Recently, dependent models have gained significant attention in machine learning, particularly for modeling diverse, structured, and uncertain data. The interactions between points often encode notions of repulsion or diversity, which are valuable in machine learning for improving generalization and reducing redundancy. The aim of this thesis is to develop a comprehensive theory of strongly dependent models with applications in machine learning.

Author dc:creator

• TRAN HOANG SON

• Machine Learning
• Statistical Physics
• Probability
• Interacting particle systems

Licence

• https://scholarbank.nus.edu.sg/bitstreams/e0ff69ee-19f2-48f1-93a9-6bd72e7b6c41/download