Faezeh Habibi
Faezeh Habibi is an AI Researcher at SingularityNET and a Doctoral Student in Computational Neuroscience at the Neural Adaptive Computing (NAC) Laboratory at the Rochester Institute of Technology (RIT), where she is supervised by Professor Alexander G. Ororbia II.
Faezeh specializes in predictive coding, causal learning, dynamical systems, and representation learning and contributes to the development of biologically inspired alternatives to backpropagation for neural network training. Her research focuses on advancing predictive coding as a promising framework toward Artificial General Intelligence (AGI).
At SingularityNET, Faezeh conducts research on predictive coding and its applications to decentralized AI, working under the broader vision of Dr. Ben Goertzel and the Artificial Superintelligence (ASI) Alliance. She has contributed significantly to the Neural Generative Coding (NGC) framework developed by the NAC Lab, and is actively involved in the ActPCGeom project, which explores the application of information geometry metrics, specifically Fisher-Rao and Wasserstein, to accelerate and improve neural learning performance within predictive coding models.
Faezeh is a featured speaker at major AI conferences, presenting on the role of predictive coding on the path to AGI. At the AGI-25 Conference, the 18th Annual AGI Conference held at Reykjavík University in Iceland in August 2025, she discussed why current AI models cannot capture the characteristics required for general intelligence.
She also presented at the Beneficial AGI Summit (BGI 2025) in Istanbul, Turkey, where her talk on neural networks trained by predictive coding was recognized as one of the summit’s standout presentations. Her work highlights how predictive coding offers architectures designed for transparency, controllability, and epistemic humility, representing a safer alternative to current dominant AI development approaches.
She also presented at a DeAGI & Cocktails event hosted by the ASI Alliance, where she walked through predictive coding as a promising alternative to backpropagation for training neural networks. Read VIDI: A Video Dataset of Incidents and A branch and bound technique for finding the minimal solutions of the linear optimization problems subjected to Lukasiewicz.
Before joining SingularityNET in March 2025, Faezeh gained international research experience across multiple leading institutions. Between October 2022 and July 2023, she was a Research Assistant in the Department of Computer Science at Aalto University in Finland, having initially joined as a Summer Research Intern in July 2022.
From February to July 2022, she completed a Research Internship at the Max Planck Institute for Dynamics of Complex Technical Systems in Germany, focusing on dynamical systems. Between October 2021 and February 2022, she undertook a Research Internship at Istanbul Technical University, where she worked on the VIDI++ project, expanding the proposed video incident detection dataset with class-negative samples under the supervision of Professor Hazım Kemal Ekenel. From July to October 2021, she completed a Research Internship at the Singapore University of Technology and Design (SUTD).
Faezeh also served as a Teaching Assistant at the College of Engineering, University of Tehran, teaching Statistics and Probability under Professor S. Mahmoud Taheri from September to December 2021 and Linear Algebra from February to July 2021.
Faezeh is currently pursuing her PhD in Computational Neuroscience at the Rochester Institute of Technology in New York, USA, which she began in August 2023. She earned her Bachelor’s Degree of Science in Engineering Science with a Minor in Graph Theory in 2023 from the University of Tehran.
She attended Farzanegan 5, a selective high school administered by the National Organization for the Development of Exceptional Talents (NODET), where she earned her High School Diploma in Mathematics. NODET is Iran’s premier educational organization for gifted students, whose notable alumni include the late Fields Medal recipient Maryam Mirzakhani.
Visit her LinkedIn profile, Google Scholar page, Personal Homepage, and GitHub page.