Selected Publications
Grammar as Control: Modular Language Generation for the Long Tail
ACL 2026
🚀 This work enabled our first MT systems from English into four languages: Kwanyama, Oshindonga, Rukwangali, and Herero. None of these are supported in commercial MT systems.
Sentiment Analysis and Language Models for Kwanyama
LREC 2026
1B, 3B, and 8B parameter LLMs for Kwanyama
On Linearizing Structured Data in Encoder-Decoder Language Models: Insights from Text-to-SQL
NAACL 2024
A Gradually Soft Multi-Task and Data-Augmented Approach to Medical Question Understanding
ACL 2021
PATTY: A Taxonomy of Relational Patterns with Semantic Types 🏆 Best Paper Runner-Up
EMNLP 2012