I am Jay Rothenberger, a Graduate Research Assistant at the University of Oklahoma (OU) studying Computer Science under Dimitris Diochnos. I study semi-supervised learning particulary for computer vision and I am interested in more data and compute efficient methods for deep learning. I believe that the future is closer to the edge than it is to HPC. I have spent two summers in industry performing machine learning research for the atmosphere with the goal of producing trustworth models. Currently I am funded under the NSF Trustworthy AI Instutute for Weather, Climate, and Coastal Oceanography. I am open to opportunities for research collaboration, so do not hesitate to contact me.
My Best Work
Meta Co-Training
GitHub: https://github.com/JayRothenberger/Meta-Co-Training
ArXiv paper: https://arxiv.org/abs/2311.18083
In many practical computer vision scenarios unlabeled data is plentiful, but labels are scarce and difficult to obtain. As a result, semi-supervised learning which leverages unlabeled data to boost the performance of supervised classifiers have received significant attention in recent literature. One major class of semi-supervised algorithms is co-training. In co-training two different models leverage different independent and sufficient "views" of the data to jointly make better predictions. During co-training each model creates pseudo labels on unlabeled points which are used to improve the other model. We show that in the common case when independent views are not available we can construct such views inexpensively using pre-trained models. Co-training on the constructed views yields a performance improvement over any of the individual views we construct and performance comparable with recent approaches in semi-supervised learning, but has some undesirable properties. To alleviate the issues present with co-training we present Meta Co-Training which is an extension of the successful Meta Pseudo Labels approach to two views. Our method achieves new state-of-the-art performance on ImageNet-10% with very few training resources, as well as outperforming prior semi-supervised work on several other fine-grained image classification datasets.
Contact
email: jay.c.rothenberger@ou.edu
GitHub: https://github.com/JayRothenberger
Linkedin: https://www.linkedin.com/in/jay-rothenberger/