Information Theory & Causal Inference
Transfer entropy, mutual information, causal discovery, uncertainty quantification, interpretable AI, and information flow analysis in complex systems.
Information Theory • Statistical Signal Processing • Causal AI
Deniz Gencaga develops information-theoretic and statistical methodologies for discovering causal structure, quantifying uncertainty, and modeling complex dynamical systems. His research combines transfer entropy, mutual information, Bayesian inference, machine learning, and signal processing, with applications ranging from speech and audio intelligence to climate dynamics, robotics, and generative AI.
Special Faculty Member, Carnegie Mellon University
Assistant Professor, Antalya Bilim University
My research focuses on developing information-theoretic, statistical, and machine learning methodologies for uncovering structure, causality, and information flow in complex dynamical systems. A central theme of my work is the use of transfer entropy, mutual information, and related measures to move beyond correlation and quantify directional interactions, emergent behavior, and predictive relationships in high-dimensional data. These methods have been applied across diverse domains including speech and audio intelligence, neuroscience, climate dynamics, remote sensing, biomedical signals, and industrial monitoring. By combining rigorous mathematical modeling with data-driven learning approaches, my research aims to advance the understanding, prediction, and control of complex natural and engineered systems.
Research Program
Transfer entropy, mutual information, causal discovery, uncertainty quantification, interpretable AI, and information flow analysis in complex systems.
Speech processing, speaker characterization, audio watermarking, speech biometrics, deep generative models, and multimodal AI.
Bayesian inference, particle filtering, Kalman methods, heavy-tailed modeling, time-series analysis, and stochastic dynamical systems.
Climate dynamics, Earth-system interactions, remote sensing, nonlinear dependencies, teleconnections, and information-theoretic analysis of environmental data.
Selected Projects
Development of advanced signal processing techniques for perceptually transparent and robust audio watermarking, including amplitude, phase, spread-spectrum, and feature-based strategies across time and frequency domains.
Transfer entropy and mutual information methods for identifying nonlinear cause-effect dependencies among climate variables, including applications to cloud coverage, sea surface temperature, and aerosol optical depth analysis.
Voice biometrics, forensic anthropometry from speech, voice disguise analysis, and computer vision methods for mission-readiness evaluation.
Statistical design of experiments, process analytics, anomaly detection, industrial data analysis, and patented contributions to materials and manufacturing research.
Publications
Teaching
Introduction to Generative AI, machine learning foundations, probability, stochastic processes, and information theory.
Digital signal processing, statistical signal processing, speech processing, and time-series analysis.
Feedback and control systems, circuit theory, robotics, telecommunications, and laboratory-based engineering education.
Talks, Awards, and Service
Keynote speaker, ACDSA International Conference on Artificial Intelligence, Computer, Data Sciences and Applications.
Best Paper Award, International Workshop on Biometrics and Forensics.
Senior Member of IEEE.
NATO-TUBITAK Research Fellowship, Consiglio Nazionale delle Ricerche, Italy.
Contact
For research collaboration, invited talks, student advising, and academic communication, please use the contact links below.
Email: d.gencaga@ieee.org
CMU Email: denizg@andrew.cmu.edu
Location: Pittsburgh, PA / Antalya, Turkey