Home

Hello there, I'm İbrahim Ethem

I'm a PhD student in Cognitive Science at METU, Ankara, working under the supervision of Dr. Duygu Ataman.

Before this, I did my BSc in Computer Engineering at Koç University and my MSc in Cognitive Science here at METU, where I worked on semantic parsing and compositional generalization in transformer models.

PhD Research

I work on reasoning in large language models, specifically on whether and how they can be made to reason better, not just perform better on benchmarks. My approach draws on interactionist theories of reasoning from cognitive science and connects them to reinforcement learning and argumentation. The broader question underneath all of it is what it would even mean for a computational model to reason.

Master’s Thesis

Transformer Models for Translating Natural Language Sentences into Formal Logical Expressions
Supervised by Assoc. Prof. Dr. Aziz Zambak and Dr. Ceyhan Temürcü.

I tested whether transformer-based models can perform semantic parsing into first-order logic, treating it as a probe for compositional generalization: can a connectionist system be truly productive, or does it just interpolate? A key contribution was a set of task-specific metrics that disentangle lexical, syntactic, and semantic error sources, making evaluation much more informative than standard accuracy scores.

Full thesis (METU Open Archive) · Pre-proceedings (PhilArchive)

Research Interests

Beyond my doctoral work, I am interested in logic and argumentation theory, the philosophy of AI and mind, and the philosophy of mathematics and language.

Contact

Mail · GitHub · Google Scholar · Academia