RPI alumnus Burak Varıcı, who earned his Ph.D. in Electrical, Computer, and Systems Engineering in 2024, has been named a recipient of the 2025 IEEE Signal Processing Society Best Ph.D. Dissertation Award. This prestigious honor recognizes doctoral research that demonstrates exceptional scientific impact, originality, clarity of presentation, timeliness, and relevance, advancing both the theory and practice of signal processing.
Varıcı’s dissertation, “Causal Learning via Interventions: Estimation and Design,” spans a broad range of topics in causal inference and learning, with a particularly significant contribution to the rapidly growing field of causal representation learning. While deep learning has achieved extraordinary empirical success, it has also highlighted a fundamental limitation: opacity. Conventional representation learning models often operate as powerful but inscrutable black boxes, offering little insight into the underlying factors and mechanisms that generate observed data.
To address this challenge, Varıcı’s work focuses on developing methods that disentangle latent generative factors in ways that are meaningfully aligned with the true data-generating process. When successful, these approaches transform learned representations from opaque objects into structured, interpretable, and scientifically actionable abstractions. This framework has wide-ranging applications across domains, including robotics, genomics, blind source separation, medical imaging, and explainable artificial intelligence.
During his time at RPI, Varıcı was advised by Ali Tajer, Professor of Electrical, Computer, and Systems Engineering. “Burak has made remarkable contributions by advancing a new vision for causal representation learning and developing novel theoretical tools that enable the systematic design of algorithms across a wide range of settings and applications,” says Tajer. “I believe his approaches will be highly impactful for many years to come.”
As a postdoctoral researcher in the Machine Learning Department at Carnegie Mellon University, Varıcı’s current research focuses on identifiable representation learning to explain and improve modern self-supervised pretraining, yielding representations that are more efficient, interpretable, and transferable. As a student at RPI, he was also a recipient of the Allen B. Dumont Prize for his doctoral dissertation, an IBM Fellowship, a Belsky Award for RPI Computational Sciences and Engineering, and a Jerry Dziuba ECSE Graduate Student Service Award.
Varıcı will receive one of the IEEE Signal Processing Society’s most prestigious awards at ICASSP 2026, to be held in Barcelona, Spain.