Daniel Lowd
 |
E-mail: lowd at uoregon dot edu
Mailing Address: |
I am an Associate Professor in the Department of Computer and Information Science at the University of Oregon.
My research interests include learning and inference with probabilistic graphical models, adversarial machine learning, and statistical relational machine learning.
You can also find me on BlueSky (@lowd.bsky.social), Twitter (@dlowd) and Mastodon (@lowd@sigmoid.social).
PROSPECTIVE STUDENTS
Interested in doing machine learning research at the University of Oregon? Please check out the FAQ for more information, then email me if you still have questions.NEWS
February 2023
SaTML 2023 paper (with Zayd Hammoudeh) on robust regression: Reducing Certified Regression to Certified Classification for General Poisoning Attacks.
JMLR paper (with Jonathan Brophy and Zayd Hammoudeh) on influence estimation for tree ensembles: Adapting and Evaluating Influence-Estimation Methods for Gradient-Boosted Decision Trees.
November 2020
NeurIPS 2022 paper (with Jonathan Brophy) on estimating probabilities using tree ensembles: Instance-Based Uncertainty Estimation for Gradient-Boosted Regression Trees.
November 2020
CCS 2022 paper (with Zayd Hammoudeh) on defending against training set attacks: Identifying a Training-Set Attack's Target Using Renormalized Influence Estimation.
November 2020
ICML 2021 paper (with Jonathan Brophy) on data removal for random forests: Machine unlearning for random forests (Supplement).
November 2020
NeurIPS 2020 paper (with Zayd Hammoudeh) on positive-unlabeled learning: Learning from Positive and Unlabeled Data with Arbitrary Positive Shift. Code: https://github.com/ZaydH/arbitrary_pu
July 2019
New CACM paper on Markov logic: Unifying Logical and Statistical AI with Markov Logic.
April 2019
I spoke at the Sisters Science Club in Sisters, Oregon on the topic "Can Artificial Intelligence Fight Alternative Facts?" (Slides)
March 2019
I'm co-organizing the 3rd Workshop on Tractable Probabilistic Modeling, co-located with ICML 2019.
January 2019
I gave a QuackChat pub talk titled, "When Can We Trust Artificial Intelligence?" (Slides)
July 2018
Our work on adversarial examples for machine translation was featured in an article in The Register.
June 2018
Our new COLING 2018 paper was selected as an Area Chair Favorite: On Adversarial Examples for Character-Level Neural Machine Translation (with Javid Ebrahimi and and Dejing Dou). (Source code)
In the news: I wrote an general audience article for The Conversation -- "Can Facebook use AI to fight online abuse?". It's been reprinted by Scientific American, Salon, The Seattle P.I., and many other outlets!
May 2018
I am co-organizing the ICML/IJCAI 2018 Workshop on Tractable Probabilistic Models (Deadline: May 18th, 2018)
New paper accepted to ACL 2018: HotFlip: White-Box Adversarial Examples for Text Classification (with Javid Ebrahimi, Anyi Rao, and Dejing Dou).
May 2017
I was awarded tenure! I'm now an Associate Professor (effective September 2017).
April 2017
I was interviewed for a BBC article on adversarial machine learning, "How to Fool Artificial Intelligence".
March 2017
Pedram Rooshenas successfully defended his dissertation and completed his Ph.D. Congratulations, Dr. Rooshenas!
February 2017
Our work on Collective Classification of Social Network Spam (with Jonathan Brophy) has been published in the 2017 AAAI Workshop on Artificial Intelligence and Cyber-Security (AICS).
October-December 2016
Two new publications on stance classification in tweets (with Javid Ebrahimi and Dejing Dou): Weakly Supervised Tweet Stance Classification by Relational Bootstrapping (COLING'16) and A Joint Sentiment-Target-Stance Model for Stance Classification in Tweets (EMNLP'16).
July 2016
Ali Torkamani successfully defended his dissertation and completed his Ph.D. Congratulations, Dr. Torkamani!
I gave an invited talk on "Adversarial Statistical Relational AI" at the IJCAI 2016 Workshop on Statistical Relational AI (StarAI). (Slides)
June 2016
We've received funding from the DARPA Media Forensics (MediFor) program! Over the next 4 years, we will develop Markov logic networks and algorithms to reason about fraudulent images and videos. (Joint team with SRI Princeton and NYU.)
May 2016
Our paper Discriminative Structure Learning of Arithmetic Circuits was published in the 19th International Conference on Artificial Intelligence and Statistics (AISTATS) (with Pedram Rooshenas).
February 2016
Our paper A Probabilistic Approach to Knowledge Translation (with Shangpu Jiang and Dejing Dou) was published in AAAI 2016
January 2016
The Libra Toolkit for Probabilistic Models has been published in the JMLR Open Source Software track! (with Pedram Rooshenas)