Independent Research Fellow
Published:
- Designed and implemented a Multi-View Canonical Correlation Analysis framework to align semantic representations across five languages for cross-lingual fake news detection
- Developed adaptive correlation-based weighting and ensemble classification pipelines, achieving up to 7.4% accuracy and 7.7% F1-score improvements over single-language baselines on the TALLIP-FakeNews dataset
- Paper published in the MIT Undergraduate Research Journal (MURJ)