About me

Hello, I am a Computer Science PhD student at Johns Hopkins University CLSP working on Natural Language Processing, advised by Prof. Mark Dredze. Before joining JHU, I was a student at Yale University, where I first learned and researched NLP under the guidance of Prof. Arman Cohan. I have broad interests in machine learning, natural language processing, and applied mathematics and statistics, especially in health, social, and business applications. I regard many practical problems as an optimization problem with partially observed data. How to make the most use of the limited data and how to maximize the quantified outcome through simple, elegant ideas and rigorous definitions are my primary focuses during research.

My current research interests include (take health applications for example):

  • Information Extraction and Affective Computing: extract user’s current physical and mental state, personality, ideology, preferences, and interests through texts, signals, and images.

  • Representation Learning: user/patient dynamic representations through longitudinal, heterogeneous data (e.g. Electronic Health Records, Social Networks, Wearables) and the extracted information.

  • Personalized Treatment: given the resources (action space) and observations (state space), optimize the treatment planning to maximize clinical outcome metrics, with a trade-off between effectiveness and [side-effects, costs, etc.].
  • Empathetic Chatbot: understand the empathy mechanism and improve chatbot’s empathetic capacity.
  • Text Generation: summarization and simplification of long complex documents, (e.g. medical/law texts) to promote accessibility for laypeople and reading disorder people.
  • Explainable AI: understand the fairness and bias of models to prevent harm in real-world applications; provide insights into the underlying phenomenons and model behaviors;

To scale up affordable and accessible healthcare services in mental therapy, I envision one day chatbots can act at a trained clinical expert level to understand, comfort, and guide patients to recover.

Publications

Medical Text Simplification: Optimizing for Readability with Unlikelihood Training and Reranked Beam Search Decoding

Lorenzo Jaime Yu Flores, Heyuan Huang, Kejian Shi, Sophie Chheang, and Arman Cohan

EMNLP 2023 Findings / code

Teaching

Teaching Assistant in CPSC 477/577 Natural Language Processing, Yale University, Jan/2024 - May/2024

Miscellaneous

I love to help people take photos and my aesthetic composition always gives them happiness and touching feelings during traveling and daily life.

My parents raised a dog during my undergraduate time. She is 5 years old now and has a similar personality to mine when I was a kid… 🤨

My favorite book and its adapted musical is Les Misérables by Victor Hugo.