Berkay Bayramoğlu

Undergraduate Student Researcher | Department of Software Engineering, Ankara University
Berkay Bayramoğlu

Biography

Berkay Bayramoğlu is an AI researcher focusing on deep learning, computer vision, and large language models (LLMs). His work addresses fundamental challenges in representation learning, structural inductive biases, and model robustness. In this context, he designs graph-augmented vision transformers, probes internal representations and uncertainty dynamics in language models, and develops relational neural frameworks for complex structured reasoning. His academic interests lie at the intersection of computer vision, geometric intelligence, and natural language processing, with an emphasis on designing mathematically grounded neural architectures for complex perceptual and reasoning tasks.

Research Interests

Computer Vision Geometric Intelligence Deep Learning Large Language Models Natural Language Processing Graph Theory Quantum Computing

Education

  • B.Sc., Software Engineering — Ankara University, 2028 (Expected)

Selected Publications

  • B. Bayramoglu, I. Erol, R. S. Kiziltepe, M. Karakus (2026) “Synthetic Image Detection under Distribution Shift: CLIP-FFT Fusion and ForensicGAT” — MediaEval 2026 Workshop, Amsterdam Published Distinctive Mention Award
  • B. Bayramoglu, M. Ficici, F. Bildirici, R. S. Kiziltepe, M. Karakus (2026) “A White-Box Benchmark for Intrinsic Hallucination Detection in Turkish LLMs” — 2026 5th International Informatics and Software Engineering Conference (IISEC), IEEE Xplore Published
  • C. S. Akdag, B. Bayramoglu, E. Celik, E. Dasci, M. Ficici, R. S. Kiziltepe, M. Karakus (2026) “Hybrid Deep Learning for 5G Positioning: Integrating Map-Derived Environmental Features with GAT-GRU-Attention” — 2026 9th International Balkan Conference on Communications and Networking (BalkanCom), IEEE Xplore Published
  • B. Bayramoglu, I. Erol, S. Akhtar, M. Karakus, R. S. Kiziltepe (2026) “Decoding Memory at the Neural Level: wPLI Connectivity and Quantum Tensor Networks for EEG-Based Memorability Prediction” — MediaEval 2026 Workshop, Amsterdam Published
View full list on Google Scholar →

Service & Talks

  • Session organizer: EXIT-ANA 2026 Special Session (ICHORA 2026)
  • Tutorials: 34th IEEE SIU 2026, ICHORA 2026, ICFNDS 2026
  • Projects: Led several TÜBİTAK-funded and industry-supported research projects
  • Honors: MediaEval 2026 Distinctive Mention Award, Certificate of Appreciation for Major Event Organization (Ankara University)

Contact

  • berkaybbayramoglu@gmail.com
  • Department of Software Engineering, Ankara University, Ankara, Turkey