Dr. Uri Shaham

Machine Learning, Representation Learning, and Causal Inference

When a streaming platform attempts to recommend what to watch from thousands of movies and series, the challenge is to understand what will truly interest each user—even when their personal taste is complex or based on limited information. To achieve this, the system must identify hidden patterns within vast amounts of data.

Dr. Uri Shaham researches how to teach systems to understand what matters within complex data, identify meaningful connections, and make more accurate decisions. His primary research focus is Representation Learning—developing methods that teach computers to organize information so that its underlying meaning becomes clear.

His research encompasses Deep Learning, Multi-modal Learning, Reinforcement Learning, and Causal Inference. Currently, he is developing systems that connect text, image, and sound even without direct alignment between them, as well as systems that learn not only what works—but also why.

Key Research Areas:

• Representation Learning

• Deep Learning

• Multi-modal Learning

• Reinforcement Learning

• Causal Inference

• Generalizable Spectral Methods

Research Nature: 

Theoretical-Applied: Strong mathematical foundations alongside the development of novel methods for Machine Learning and Deep Learning.

Career Paths:

 Graduates of the research group pursue diverse roles across several sectors:

• Industry: Machine Learning Researchers, Data Scientists, and Algorithm Developers.

• Academia: Researchers and lecturers in Artificial Intelligence, Statistics, and Data Science.

• Entrepreneurship: Founders of data-driven and AI-based ventures.