I am a research assistant and PhD candidate working on deep reinforcement learning, robot learning, and contact-rich robotic manipulation.

My work connects autonomous systems, real-world robot learning, and applied machine learning for biological datasets.

Research Interests

View research
  • Deep Reinforcement Learning
  • Robot Learning
  • Contact-rich Manipulation
  • World Models
  • Autonomous Systems
  • Machine Learning for Biological Data

Selected Publications

View all
2025
PublishedSelected

Enhancing crayfish sex identification with Kolmogorov-Arnold networks and stacked autoencoders

Yasin Atılkan, Berk Kirik, Eren Tuna Acikbas, Fatih Ekinci, Koray Acici, Tunc Asuroglu, Recep Benzer, Mehmet Serdar Güzel, Semra Benzer

Scientific Reports, 16, Article 3971

DOI: 10.1038/s41598-025-34095-z

This study investigates crayfish sex identification using traditional machine learning, deep learning, Kolmogorov-Arnold networks, and autoencoder-based feature extraction. It highlights the potential of hybrid artificial intelligence models for biological classification problems.

  • Machine Learning
  • Deep Learning
  • Kolmogorov-Arnold Networks
  • Autoencoders
2024
PublishedSelected

Advancing crayfish disease detection: A comparative study of deep learning and canonical machine learning techniques

Yasin Atılkan, Berk Kirik, Koray Acici, Recep Benzer, Fatih Ekinci, Mehmet Serdar Güzel, Semra Benzer, Tunc Asuroglu

Applied Sciences, 14(14), 6211

DOI: 10.3390/app14146211

This study compares deep learning and canonical machine learning techniques for crayfish disease detection using an imbalanced dataset. It explores hybrid approaches that combine deep feature extraction with traditional classifiers.

  • Machine Learning
  • Deep Learning
  • Disease Detection
  • Crayfish
2024
PublishedSelectedTurkish

Otonom araçlarda derin pekiştirmeli öğrenme yöntemleri ile sollama

Overtaking with deep reinforcement learning methods in autonomous vehicles

Fehim Köylü, Yasin Atılkan

Niğde Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi, 13(2), 429-439

DOI: 10.28948/ngumuh.1331354

This article studies the autonomous overtaking problem using deep reinforcement learning methods, including curriculum learning-based training strategies for autonomous vehicle decision-making.

  • Deep Reinforcement Learning
  • Autonomous Vehicles
  • Curriculum Learning
  • Overtaking

Featured Projects

View all
ActiveFeatured

Contact-rich Robotic Manipulation with Deep Reinforcement Learning

Ongoing PhD research titled Deep Reinforcement Learning in Contact-Rich Robotic Manipulation Tasks, with an emphasis on insertion and assembly-style robot learning problems.

  • Deep Reinforcement Learning
  • Robot Learning
  • Contact-rich Manipulation
  • Peg-in-hole
CompletedFeatured

Autonomous Overtaking with Deep Reinforcement Learning

Master's thesis and journal publication work on autonomous overtaking as a deep reinforcement learning decision-making problem for autonomous systems.

  • Deep Reinforcement Learning
  • Autonomous Driving
  • Autonomous Overtaking
  • Curriculum Learning
  • PPO
  • DQN
  • A2C