ENG-DRB is a discourse relation dataset built specifically for engineering content, introduced in a recent
paper at IJCNLP-AACL 2025.
The dataset analyzes over 1,200 relations extracted from real tutorial videos across civil, mechanical, and electrical engineering and shows that even advanced AI models struggle to accurately understand engineering discourse, particularly implicit causal and temporal relations. This work highlights a gap between general-purpose NLP and the specialized language used by engineers. The
dataset and code are open-source and publicly available.
Jan 27, 2026
We are pleased to share that Prof. Pingbo Tang has been appointed as a Director of the Engineers’ Society of Western Pennsylvania (ESWP).
Jan 23, 2026
Our work on human–AI collaboration for safe and efficient production of building mechanical system components and modular housing units was featured on the Autodesk Research Blog. The project is supported by the Pennsylvania Department of Community & Economic Development, with industry partners DMI Companies, Inc. and Module Design, and collaborators Burcu Akinci, Yorie Nakahira, Ruoxin Xiong, Xiyu Deng, Miaosi Dong, Joonsun Hwang, Yanyu Wang, Haoming Jing, and Yael Netser.
Nov 19, 2025
Safety21 researcher Prof. Pingbo Tang was selected for a 2025 Engineering faculty award. He received the Steven J. Fenves Award for Systems Research.
Sep 29, 2025
Seongeun Park Receives the CUAHSI Hydroinformatics Innovation Fellowship
Seongeun was selected for the CUAHSI Hydroinformatics Innovation Fellowship, supported by NSF. Her project applies AI methods to strengthen training and decision-making in remote infrastructure operations through virtual AI avatars.
Aug 18, 2025
Our research group is developing AI-powered frameworks to support the safe and efficient deployment of nuclear microreactors. The approach uses large language models and knowledge graphs to align procedures, regulations, and training across different countries.
Jul 14, 2025
