A two-stage deep learning pipeline that automatically classifies news articles into 6 categories and detects 12 AI ethics issues using TF-IDF, RoBERTa, and ONNX optimization. Achieves 81.7% F1-score on category classification and 79.8% on multi-label ethics detection across 26K+ articles. All the data is scraped from different portals.

A robust news article classification system combining a two-stage editorial bias modeling strategy with transformer-based ensemble methods under noisy, real-world data conditions.

Flask tabanlı, İki Aşamalı Doğrulama (Two-Stage Validation) mekanizmasına sahip, Derin Öğrenme (Deep Learning) ile Beyin MR’ları üzerinden tümör tespiti ve güvenilirliği artırma projesidir.