Bio
I am a Senior Scientist at Bloomberg. I obtained my PhD from Johns Hopkins University, where I was advised by Prof. Benjamin Van Durme. My research interests are structured prediction (e.g. semantic parsing, code generation, information extraction) and sequence modeling (e.g. machine translation, streaming sequence transduction).
Recently, I have been working on large language models (LLMs) projects including using reinforcement learning (RL) to improve model steerability and robustness, retrieval augmented generation (RAG) for long-form question answering, and model fine-tuning with preference and demonstration data. I was awarded outstanding paper for Learning to Retrieve Iteratively for In-Context Learning at EMNLP 2024 and Iterative Document-level Information Extraction via Imitation Learning at EACL 2023.
Preprints
Streaming Sequence Transduction through Dynamic Compression
Publications
Learning to Retrieve Iteratively for In-Context Learning
FaithScore: Evaluating Hallucinations in Large Vision-Language Models
The Language Barrier: Dissecting Safety Challenges of LLMs in Multilingual Contexts
Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation
Narrowing the Gap between Zero- and Few-shot Machine Translation by Matching Styles
MultiMUC: Multilingual Template Filling on MUC-4
A Unified View of Evaluation Metrics for Structured Prediction
Condensing Multilingual Knowledge with Lightweight Language-Specific Modules
On Event Individuation for Document-Level Information Extraction
When Do Decompositions Help for Machine Reading?
Iterative Document-level Information Extraction via Imitation Learning
Differentiable Tree Operations Promote Compositional Generalization
An Empirical Study on Finding Spans
Everything Is All It Takes: A Multipronged Strategy for Zero-Shot Cross-Lingual Information Extraction
LOME: Large Ontology Multilingual Extraction
Pattern-aware Data Augmentation for Query Rewriting in Voice Assistant Systems
Hierarchical Entity Typing via Multi-level Learning to Rank
Joint Modeling of Arguments for Event Understanding
Reading the Manual: Event Extraction as Definition Comprehension