Hi, my name is Denny! I’m a Master’s student in Computer Science at Tufts University, where I’m advised by Prof. Michael Hughes and Prof. Jivko Sinapov.

I am broadly interested in deep learning with constraint. That constraint could be a lack of labeled data, quality data, rich data representations, or inherent model limitations.

To address these constraints, I have explored diverse approaches including:

• Transfer learning
• Weakly-supervised methods like multiple instance learning
• Reinforcement learning for LLM post-training
• Parameter-efficient fine-tuning with LoRA
• Simulation-to-real transfer
• Incorporating inductive biases inspired by cognitive science.

I am also interested in deep learning interpretability because as these models become ubiquitous in daily life, understanding their decision-making is essential for safe deployment.

My previous work spans medical imaging, computer vision for robotics, vision-language models, and computational chemistry.

Open Source Contributions

Recent News

October 2025 - Our paper “Synthetic Data Reveals Generalization Gaps in Correlated Multiple Instance Learning” has been accepted at the ML4H Findings Track! Read the paper on arXiv

I’ll be in San Diego for the ML4H conference and staying for NeurIPS. If you’d like to chat, let me know! ☕

July 2024 - Featured in CMU ChemE Alumni Spotlight!