9 papers
Online Label-Free Gradient Matching for Continual Language Model Training
Sajjad Pakdamansavoji, Mohammad Reza Rezaei, Ilija Subasic +3
Language models in customer-facing products must be continually updated, since the data they serve shifts over time. As retraining on the full accumulated data is impractical, each…
Fragility Is Not Reducible to Hardness: Conditional Causal Difficulty for LLMs
Mohammad Reza Rezaei, Dingxian Wang, Nikhil Verma +3
Large language models show large and seemingly arbitrary performance drops when a task is rewritten in ways that leave its requirements untouched. We recast task difficulty as a ca…
How VLAs (Really) Work In Open-World Environments
Amir Rasouli, Yangzheng Wu, Zhiyuan Li +4
Vision-language-action models (VLAs) have been extensively used in robotics applications, achieving great success in various manipulation problems. More recently, VLAs have been us…
Do World Action Models Generalize Better than VLAs? A Robustness Study
Zhanguang Zhang, Zhiyuan Li, Behnam Rahmati +11
Robot action planning in the real world is challenging as it requires not only understanding the current state of the environment but also predicting how it will evolve in response…
Distracted Robot: How Visual Clutter Undermine Robotic Manipulation
Amir Rasouli, Montgomery Alban, Sajjad Pakdamansavoji +4
In this work, we propose an evaluation protocol for examining the performance of robotic manipulation policies in cluttered scenes. Contrary to prior works, we approach evaluation…
Improving Robotic Manipulation Robustness via NICE Scene Surgery
Sajjad Pakdamansavoji, Mozhgan Pourkeshavarz, Adam Sigal +3
Learning robust visuomotor policies for robotic manipulation remains a challenge in real-world settings, where visual distractors can significantly degrade performance and safety.…
WALDO: Where Unseen Model-based 6D Pose Estimation Meets Occlusion
Sajjad Pakdamansavoji, Yintao Ma, Amir Rasouli +1
Accurate 6D object pose estimation is vital for robotics, augmented reality, and scene understanding. For seen objects, high accuracy is often attainable via per-object fine-tuning…
Ground Plane Projection for Improved Traffic Analytics at Intersections
Sajjad Pakdamansavoji, Kumar Vaibhav Jha, Baher Abdulhai +1
Accurate turning movement counts at intersections are important for signal control, traffic management and urban planning. Computer vision systems for automatic turning movement co…
AnyBox: Efficient Zero-Shot 9DoF Pose Estimation of Boxes for Robotic Manipulation
Yintao Ma, Sajjad Pakdamansavoji, Charles Eret +5
Recovering the 9D pose of objects, both their 6D pose and 3D dimensions, under clutter and occlusion is a core requirement for warehouse automation, logistics, and manufacturing. M…