Nahıl AutoML
End-to-end AutoML approach that automates data prep, weight init, architecture, activation & loss design, and hyperparameters.
Project
Details
Projects, Publications & Activities
Quick summaries with visuals. Click any card to read the details.
Projects
Applied and research projects
Early-Stage Breast Cancer Diagnosis (AI)
Ministry of Health & TÜSEB funded. ~1,000 expert-annotated mammograms with BI-RADS classification for mass & calcification detection at 90%+ accuracy.
Project
Details
Campus Entry/Exit AI Control
People/vehicle counting and ID tracking across six gates with real-time processing of camera streams.
Project
Details
Cafe & Restaurant Recommender (ML)
TÜBİTAK-backed project using customer segmentation and collaborative filtering for user–item–venue recommendations.
Project
Details
ROS + OpenCV Lane Detection
Real-time lane detection with a USB camera plus PID control outputs (Teknofest Robotaxi).
Project
Details
Publications
Papers and in-progress manuscripts
Automatic Design of Deep Network Activation Functions with Genetic Programming
Evolutionary search for activation functions
In journal
Details
A Comparative Review of EA & Mutation Types in Automatic Activation Function Design
Comparative look at evolutionary algorithms for activation design
Drafting
Details
Semantic Search for Automatic Activation Function with Artificial Bee Colony Programming
Semantic-guided activation search with ABC programming
Drafting
Details
Neural Architecture Search with Artificial Bee Colony Coding
New coding scheme for NAS using artificial bee colony
In development
Details
Activities
Community work, events, and talks
Erciyes University AI Club — Founder & President
20-person team; eight theoretical projects (TÜBİTAK) and a TÜSEB-backed healthcare AI project
Community
Details