This research project focuses on enhancing the trustworthiness of semantic similarity models by integrating uncertainty quantification (UQ) techniques. Traditional models compute similarity between texts using deep learning architectures like BERT or SBERT, but they often produce deterministic scores without indicating how confident the model is in its output. Our project aims to bridge this gap by building a semantic similarity system that not only measures the semantic relationship between textual inputs but also provides confidence estimates for those scores.
Raveendiran Akshiya
I'm a final-year Computer Science and Engineering undergraduate at the University of Moratuwa, passionate about software development and machine learning. Currently, I'm diving deep into semantic similarity and uncertainty quantification as part of my final-year research project. Excited to explore how these concepts can enhance the reliability and interpretability of intelligent systems.
Brief Description of Project
Recent News
- All
