Publications
A collection of my research work.
ToolScope: Enhancing LLM Agent Tool Use through Tool Merging and Context-Aware Filtering
Marianne Menglin Liu, Daniel Garcia, Fjona Parllaku, Vikas Upadhyay, Fahad Shah, Dan Roth
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) 2026
Improves LLM tool-use by automatic tool merging and context-aware tool filtering.
LAD-RAG: Layout-aware Dynamic RAG for Visually-Rich Document Understanding
Zhivar Sourati, Zheng Wang, Marianne Menglin Liu, Yazhe Hu, Mengqing Guo, Sujeeth Bharadwaj, Kyu J. Han, Tao Sheng, Sujith Ravi, Morteza Dehghani, Dan Roth
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) 2026
Layout-aware dynamic retrieval framework for multi-page visually rich document QA.
Routesplain: Towards Faithful and Intervenable Routing for Software-Related Tasks
Adam Storek, Vikas Upadhyay, Marianne Menglin Liu, Daniel W Peterson, Anshul Mittal, Sujeeth Bharadwaj, Syed Fahad Allam Shah, Dan Roth
Proceedings of the 3rd Conference on Language Modeling (COLM 2026) 2026
Concept-based LLM routing for software tasks with faithful and intervenable decision rationales.
PLSQLBench: Benchmarking LLM Systems for Executable Procedural Database Programming
Marianne Menglin Liu, Leonid Boytsov, Daniel W. Peterson, Pramuditha Perera, Rongguang Wang, Sai Ashish Somayajula, Syed Hamza Rafique, Rohit Saini, Shubham Pathak, Sujeeth Bharadwaj, Tao Sheng, Graham Horwood, Fahad Shah, Ankan Bansal, Sujith Ravi, Dan Roth
EMNLP Industry Track 2026
A benchmark for evaluating whether LLMs can write executable PL/SQL programs using execution-based tests.
SOMA-SQL: Resolving Multi-Source Ambiguity in NL-to-SQL via Synthetic Log and Execution Probing
Sai Ashish Somayajula‡, Marianne Menglin Liu‡, Chuan Lei, Fjona Parllaku, Daniel Garcia, Rongguang Wang, Syed Fahad Allam Shah, Ankan Bansal, Sujeeth Bharadwaj, Tao Sheng, others
‡ Equal contribution
arXiv preprint arXiv:2606.11424 2026
Resolves multi-source ambiguity in NL-to-SQL using synthetic logs and execution probing.
OraPlan-SQL: A Planning-Centric Framework for Complex Bilingual NL2SQL Reasoning
Marianne Menglin Liu‡, Sai Ashish Somayajula‡, Syed Fahad Allam Shah, Sujith Ravi, Dan Roth
‡ Equal contribution
International Joint Conference on Knowledge Graphs 2025
Planning-centric NL2SQL system that ranked first in the Archer NL2SQL Evaluation Challenge 2025.
RAG Makes Guardrails Unsafe? Investigating Robustness of Guardrails under RAG-style Contexts
Yining She, Daniel W Peterson, Marianne Menglin Liu, Vikas Upadhyay, Mohammad Hossein Chaghazardi, Eunsuk Kang, Dan Roth
arXiv preprint arXiv:2510.05310 2025
Studies context robustness issues in LLM guardrails when retrieval-augmented contexts are introduced.