Profile picture
Maeda
AI Researcher, CS PhD
Home
Research
CV
Back

Research Work

Conversational AI Agents and Retrieval-Augmented Generation

"Creating Conversational Datasets for Retrieval-Augmented Generation Applications is Hard: Challenges and Research Opportunities" Maeda F. Hanafi, Kshitij Fadnis, Marina Danilevsky, Sara Rosenthal, Yannis Katsis. CHI-LBW 2025. Extended Abstract
"Diagnosing and Prioritizing Issues in Automated Order-Taking Systems: A Machine-Assisted Error Discovery Approach" Maeda F. Hanafi, Frederick Reiss, Yannis Katsis, Robert Moore, Mohammad H Falakmasir, Pauline Wang, David Wood, Changchang Liu. CHI 2025. Paper Publication (Honorable Mention Award)
"Machine-Assisted Error Discovery in Conversational AI Systems" Maeda F. Hanafi, Frederick Reiss, Yannis Katsis, Robert Moore, David Wood, Mohammad H Falakmasir, Changchang Liu. CHI Late Breaking Work (CHI-LBW) 2024. Extended Abstract
Paper
Video Preview
Video

Human-in-the-loop Systems and Large Language Models

"Identifying Noise in Human-Created Datasets using Training Dynamics from Generative Models" Maeda F. Hanafi, Ishan Jindal, Yannis Katsis, Lucian Popa, Huaiyu Zhu. EMNLP Findings 2025. Paper
"A Comparative Analysis between Human-in-the-loop Systems and Large Language Models for Pattern Extraction Tasks" Maeda F. Hanafi, Yannis Katsis, Ishan Jindal, Lucian Popa.. DaSH@EMNLP 2022. Workshop Paper

InteractEva: A Simulation-based Evaluation Framework for Interactive AI Systems

"InteractEva: A Simulation-based Evaluation Framework for Interactive AI Systems" Yannis Katsis, Martín Santillán Cooper, Maeda F. Hanafi and Yunyao Li. AAAI 2022. Demo Paper
Paper
"A Simulation-based Evaluation Framework for Interactive AI Systems and Its Application" Martín Santillán Cooper, Maeda Hanafi, Yannis Katsis and Yunyao Li. IAAI 2022. Industry Track Paper
Paper

WhyFlow: Interactively Explaining Errors in Data Extraction Pipelines

"WhyFlow: Explaining Errors in Data Flows Interactively" Maeda F. Hanafi, Azza Abouzied, Marina Danilevsky, and Yunyao Li. SIGKDD. DASH 2020. Paper
Paper
Video
"FlowFix/WhyFlow: Diagnosing Incorrect Behavior in Large Data Pipelines" NYUAD Annual Research Conference 2017. Poster Presentation
"Diagnosing Incorrect Behavior in Large Data Pipelines" IBM Research Almaden, San Jose, California. Seminar and Poster Presentation

Texture: Structure Identification Over Print and PDF Documents

"A Collaborative Framework for Structure Identification over Print Documents" Maeda F. Hanafi, Miro Mannino, and Azza Abouzied. SIGMOD. HILDA 2019. Paper Publication
Paper
"Texture: Extracting Texts using User Annotations" NYUAD Annual Research Conference 2015. Poster Presentation
"Texture: Structuring Semi-Structured Documents" University of California, Berkeley. 2015. Seminar

SEER: Auto-Generating Extraction Rules from User-Highlighted Texts

"Synthesizing Extraction Rules from User Examples with SEER" Maeda F. Hanafi, Azza Abouzied, Laura Chiticariu, and Yunyao Li. SIGMOD 2017. Demo Paper
Paper
"SEER: Auto-Generating Information Extraction Rules from User-Specified Examples" Maeda F. Hanafi, Azza Abouzied, Laura Chiticariu, and Yunyao Li. CHI 2017. Paper Publication
Paper
Video
CHI Slides
"Seer: An End-to-End Automated Workflow for Information Extraction using User Highlights" North East Database Day, Massachusetts Institute of Technology. 2016. Poster Presentation
"SEER: An End-to-End Automated Workflow for Information Extraction using User Highlights" IBM Research Almaden, San Jose, California. 2015. Seminar and Poster Presentation

Services

Services

  • Served as research paper reviewer: AKBC 2021, EMNLP 2022, CHI 2022, CHI-LBW 2024, CSCW 2024, NAACL 2024, EMNLP 2024 Industry Track, CHI 2025, CHI-LBW 2025, EMNLP 2025 Industry Track, CHI 2026, EACL 2026
  • Served as primary associate chair: CHI-LBW 2024