1 code implementation • 6 Nov 2023 • Shiyang Lu, Haonan Chang, Eric Pu Jing, Abdeslam Boularias, Kostas Bekris
This work presents OVIR-3D, a straightforward yet effective method for open-vocabulary 3D object instance retrieval without using any 3D data for training.
Ranked #4 on 3D Open-Vocabulary Instance Segmentation on Replica
1 code implementation • 27 Sep 2023 • Haonan Chang, Kowndinya Boyalakuntla, Shiyang Lu, Siwei Cai, Eric Jing, Shreesh Keskar, Shijie Geng, Adeeb Abbas, Lifeng Zhou, Kostas Bekris, Abdeslam Boularias
We present an Open-Vocabulary 3D Scene Graph (OVSG), a formal framework for grounding a variety of entities, such as object instances, agents, and regions, with free-form text-based queries.
1 code implementation • 22 Sep 2023 • Xinyu Zhang, Yuting Wang, Abdeslam Boularias
DE-ViT establishes new state-of-the-art results on all benchmarks.
Ranked #1 on Few-Shot Object Detection on MS-COCO (30-shot)
1 code implementation • 25 Jul 2023 • Xinyu Zhang, Abdeslam Boularias
Our fine-tuning procedure outperforms state-of-the-art techniques for unsupervised semantic segmentation through linear probing, without the use of any labeled data.
no code implementations • 9 Apr 2023 • Shiyang Lu, Yunfu Deng, Abdeslam Boularias, Kostas Bekris
This work proposes a self-supervised learning system for segmenting rigid objects in RGB images.
1 code implementation • 30 Jan 2023 • Haonan Chang, Dhruv Metha Ramesh, Shijie Geng, Yuqiu Gan, Abdeslam Boularias
We present Mono-STAR, the first real-time 3D reconstruction system that simultaneously supports semantic fusion, fast motion tracking, non-rigid object deformation, and topological change under a unified framework.
1 code implementation • 7 Oct 2022 • Haonan Chang, Abdeslam Boularias
We present the first real-time system capable of tracking and reconstructing, individually, every visible object in a given scene, without any form of prior on the rigidness of the objects, texture existence, or object category.
no code implementations • 3 Jul 2022 • Liam Schramm, Yunfu Deng, Edgar Granados, Abdeslam Boularias
Dealing with sparse rewards is a long-standing challenge in reinforcement learning (RL).
no code implementations • 8 Mar 2022 • Junchi Liang, Bowen Wen, Kostas Bekris, Abdeslam Boularias
This work aims to learn how to perform complex robot manipulation tasks that are composed of several, consecutively executed low-level sub-tasks, given as input a few visual demonstrations of the tasks performed by a person.
1 code implementation • 26 Jun 2021 • Andrew S. Morgan, Bowen Wen, Junchi Liang, Abdeslam Boularias, Aaron M. Dollar, Kostas Bekris
Highly constrained manipulation tasks continue to be challenging for autonomous robots as they require high levels of precision, typically less than 1mm, which is often incompatible with what can be achieved by traditional perception systems.
1 code implementation • 10 Nov 2020 • Juntao Tan, Changkyu Song, Abdeslam Boularias
The triplet examples are finally used to train a siamese neural network that projects the generic visual features into a low-dimensional manifold.
1 code implementation • 9 Nov 2020 • Baichuan Huang, Shuai D. Han, Abdeslam Boularias, Jingjin Yu
The high accuracy of DIPN allows direct integration with a grasp network, yielding a robotic manipulation system capable of executing challenging clutter removal tasks while being trained in a fully self-supervised manner.
no code implementations • 4 Aug 2020 • Junchi Liang, Abdeslam Boularias
This paper introduces an algorithm for discovering implicit and delayed causal relations between events observed by a robot at arbitrary times, with the objective of improving data-efficiency and interpretability of model-based reinforcement learning (RL) techniques.
Model-based Reinforcement Learning reinforcement-learning +1
no code implementations • 28 Jun 2020 • Chaitanya Mitash, Rahul Shome, Bowen Wen, Abdeslam Boularias, Kostas Bekris
The effectiveness of the proposed approach is demonstrated by developing a robotic system that picks a previously unseen object from a table-top and places it in a constrained space.
Robotics
1 code implementation • L4DC 2020 • Avishai Sintov, Andrew Kimmel, Bowen Wen, Abdeslam Boularias, Kostas Bekris
Precise in-hand manipulation is an important skill for a robot to perform tasks in human environments.
Model-based Reinforcement Learning Reinforcement Learning (RL)
no code implementations • 21 May 2020 • Liam Schramm, Avishai Sintov, Abdeslam Boularias
We find that a major cause of these problems is that models trained on small amounts of data can have chaotic or divergent behavior in some regions.
no code implementations • L4DC 2020 • Changkyu Song, Abdeslam Boularias
This paper proposes a new method for manipulating unknown objects through a sequence of non-prehensile actions that displace an object from its initial configuration to a given goal configuration on a flat surface.
no code implementations • 11 May 2020 • Changkyu Song, Abdeslam Boularias
We propose a new technique for pushing an unknown object from an initial configuration to a goal configuration with stability constraints.
no code implementations • 11 Oct 2019 • Chaitanya Mitash, Bowen Wen, Kostas Bekris, Abdeslam Boularias
To evaluate this method, a dataset of densely packed objects with challenging setups for state-of-the-art approaches is collected.
no code implementations • 13 Mar 2019 • Changkyu Song, Abdeslam Boularias
We present a probabilistic approach for building, on the fly, 3-D models of unknown objects while being manipulated by a robot.
no code implementations • 25 Jun 2018 • Chaitanya Mitash, Abdeslam Boularias, Kostas Bekris
This work proposes an autonomous process for pose estimation that spans from data generation to scene-level reasoning and self-learning.
no code implementations • 18 Jun 2018 • Jean-Philippe Mercier, Chaitanya Mitash, Philippe Giguère, Abdeslam Boularias
We then show that the performance of the detector can be substantially improved by using a small set of weakly annotated real images, where a human provides only a list of objects present in each image without indicating the location of the objects.
no code implementations • 17 Jun 2018 • Junchi Liang, Abdeslam Boularias
This paper presents a new and efficient method for learning such representations.
no code implementations • 16 May 2018 • Chaitanya Mitash, Abdeslam Boularias, Kostas Bekris
The pointsets are then matched to congruent sets on the 3D object model to generate pose estimates.
no code implementations • 12 Apr 2018 • Shaojun Zhu, David Surovik, Kostas E. Bekris, Abdeslam Boularias
This paper aims to identify in a practical manner unknown physical parameters, such as mechanical models of actuated robot links, which are critical in dynamical robotic tasks.
no code implementations • 24 Oct 2017 • Chaitanya Mitash, Abdeslam Boularias, Kostas E. Bekris
Experimental results indicate that this process is able to quickly identify in cluttered scenes physically-consistent object poses that are significantly closer to ground truth compared to poses found by point cloud registration methods.
no code implementations • 24 Oct 2017 • Shaojun Zhu, Andrew Kimmel, Kostas E. Bekris, Abdeslam Boularias
This paper presents a method for identifying mechanical parameters of robots or objects, such as their mass and friction coefficients.
no code implementations • 22 Mar 2017 • Shaojun Zhu, Andrew Kimmel, Abdeslam Boularias
We consider the problem of a robot learning the mechanical properties of objects through physical interaction with the object, and introduce a practical, data-efficient approach for identifying the motion models of these objects.
1 code implementation • 9 Mar 2017 • Chaitanya Mitash, Kostas E. Bekris, Abdeslam Boularias
The models are placed in physically realistic poses with respect to their environment to generate a labeled synthetic dataset.
no code implementations • NeurIPS 2012 • Samory Kpotufe, Abdeslam Boularias
In regression problems over $\real^d$, the unknown function $f$ often varies more in some coordinates than in others.
no code implementations • NeurIPS 2012 • Abdeslam Boularias, Jan R. Peters, Oliver B. Kroemer
We present a new graph-based approach for incorporating domain knowledge in reinforcement learning applications.
no code implementations • NeurIPS 2010 • Abdeslam Boularias, Brahim Chaib-Draa
We consider the problem of apprenticeship learning where the examples, demonstrated by an expert, cover only a small part of a large state space.