Boston, Massachusetts, United States
• Leverage domain adaptation to successfully address Sample Selection Bias in multimedia recommendation systems. Deploy models on TikTok and WeChat datasets, perform online A/B testing, and achieve significant improvements. • Extracted automation rules and user policies from smart home apps to generate behavior graphs for security analysis. Built a semantic analyzer using Transformer-based architectures for NLP tasks, leveraging BERT for embedding generation, and fine-tuned the model with LoRA to accurately parse app descriptions and configurations.
Institute of Diversity Sciences (IDS) Fellowship Program