Bahaa Aldeeb

I am a researcher interested in mobile robot perception and machine learning. Exploring ways of leveraging weak signals to boostrap the learning of perception towards mobile manipulation piques my interest.

Currently, I work as a research staff member in the Robotics, Perception, and Manipulation Laboratory lead by Professor Karthik Desingh at the University of Minnesota, Twin-cities.

Prior to that I earned my Masters in Robotics in 2022 and a Bachelors in Computer Engineering in 2017, from University of Michigan. I also worked as a Perception Algorithm Engineer between 2018 and 2020.

I have had the forture of holding positions at various copanies including The University of Minnesota, The University of Michigan, Zenuity, Toyota, Tesla, and ZF.

news

Oct 25, 2023	Attending CoRL 2023 to present GANOCS in the OOD Workshop.
Aug 30, 2022	Excited to join RPM lab as a Robotics & Vision researcher.
May 1, 2022	Graduated with an M.S. in Robotics from UMich & joined Gatik to work on computer vision.

selected publications

Talk Through It: End User Directed Manipulation Learning

Carl Winge, Adam Imdieke, Bahaa Aldeeb, and 2 more authors

2024

HTML PDF
GANOCS: Domain Adaptation of Normalized Object Coordinate Prediction Using Generative Adversarial Training

Bahaa Aldeeb, Sahith Reddy Chada, and Karthik Desingh

In First Workshop on Out-of-Distribution Generalization in Robotics at CoRL, 2023

Abs PDF

Estimating 2D-3D correspondences has proven to be a very useful tool for category-level pose and scale estimation and robot manipulation tasks; how- ever, it is hindered by the difficulty of obtaining 3D object models and labels. Simulation reduces the burden of labeling but introduces a gap between the train- ing and operational domains. We introduce a novel architecture for integrating cross-domain data in the training of NOCS predictors (a form of 2D-3D corre- spondences). We leverage Generative-Adversarial-Networks (GANs) to avoid the need for burdensome real-data labeling by using a domain-agnostic discrimina- tor as a supervisor. This work presents results demonstrating the potential of our method.
NeRF-Frenemy: Co-Opting Adversarial Learning for Autonomy-Directed Co-Design

Stanley Lewis, Bahaa Aldeeb, Anthony Opipari, and 3 more authors

RSS Workshop-Implicit Representations for Robotic Manipulation, 2022

Abs HTML PDF

We explore the use of object models along with a pose estimation model, to automatically enform object design that fascilitate pose estimation.