Open Source Contributor
Published in Google Summer of Code (GSoC) 2024 @ HumanAI, 2024
- Used NLP models to analyze 500k Dark Web discussion points, identifying key topics and establishing 5 baseline categories.
- Enhanced the model to interpret both images and text using BERT (170 topics) and CLIP/Vision Transformer (3 topics).
- Validated clustering results with Machine Learning algorithms and LSTM, examining topic evolution and sentiment analysis over time.
- Deployed 8 predictive models on Hugging Face to forecast trends and topics identified in the analysis.
- Read more in this blog post and explore the project on the GitHub Repo.