
Data Mining Competition UNITY#14
YOGYAKARTA STATE UNIVERSITY • 2026
Developed a Siamese self-supervised learning model for post-disaster damage segmentation, achieving a 6% average improvement over the xBD paper baseline.

Recent public pushes • 26 minggu terakhir

An efficient hybrid sequence modeling system for Indonesian speech synthesis using multilingual knowledge transfer.

A ResNet-based road damage classification system combined with a Graph Neural Network to prioritize efficient road maintenance.

A Siamese self-supervised learning model for segmenting damaged areas after disasters, evaluated against the xBD paper baseline.

A low-latency Retrieval-Augmented Generation pipeline integrated into a mobile application to support customer-worker peer-to-peer interactions.
Insert coin to recruit Player 1 for brand identity, full-stack visual design, or art direction campaigns.