معرفی
Daniel Lowd is a Professor in the Department of Computer Science at the University of Oregon. He studies adversarial machine learning, probabilistic graphical models, and statistical relational AI, with applications to security and data poisoning defense.
His research has been funded by the DARPA Media Forensics (MediFor) program (2016), and he received tenure in 2017. He co-organizes the annual Workshop on Tractable Probabilistic Modeling, including at ICML and IJCAI.
- Key research areas: Probabilistic AI, Adversarial Learning, Markov Logic
- Major collaborations: Javid Ebrahimi, Jonathan Brophy, Zayd Hammoudeh
Recent publications include work on robust regression (SaTML 2023), influence estimation (JMLR 2023), and neural score estimation (ICML/NeurIPS 2021-2023). He maintains the Libra Toolkit for probabilistic models.
As an active academic, he advises PhD students like Shivvrat Arya (now at UTD), Jonathan Brophy, and Zayd Hammoudeh. He advocates for faculty unionization and open science, and is a unicyclist in his free time.
