 |
 |
|
I am a 4th year PhD student at the University of Illinois at Urbana-Champaign working with Nancy Amato on Abstraction. I completed my undergraduate degree in Math and Computer Science at Rutgers University.
I am interested in how the representations we use to plan affect decision making. In Robotics, and in AI in general, we deal with search spaces that are too large to be explored exhaustively, and so effective planning must be guided by prior knowledge which captures lower dimensional underlying structure in environments. In other words, planning should be done at a higher level, utilizing abstract state and action representations.
Publications
Discrete State-Action Abstraction via the Successor Representation, Amnon Attali, Pedro Cisneros-Velarde, Marco Morales, Nancy M. Amato, arXiv preprint, May 2023. @article{attali2022discrete, While the difficulty of reinforcement learning problems is typically related to the complexity of their state spaces, Abstraction proposes that solutions often lie in simpler underlying latent spaces. Prior works have focused on learning either a continuous or dense abstraction, or require a human to provide one. Information-dense representations capture features irrelevant for solving tasks, and continuous spaces can struggle to represent discrete objects. In this work we automatically learn a sparse discrete abstraction of the underlying environment. We do so using a simple end-to-end trainable model based on the successor representation and max-entropy regularization. We describe an algorithm to apply our model, named Discrete State-Action Abstraction (DSAA), which computes an action abstraction in the form of temporally extended actions, i.e., Options, to transition between discrete abstract states. Empirically, we demonstrate the effects of different exploration schemes on our resulting abstraction, and show that it is efficient for solving downstream tasks.
Keywords: Abstraction
Links : [ArXiv] BibTex
title={Discrete State-Action Abstraction via the Successor Representation},
author={Attali, Amnon and Cisneros-Velarde, Pedro and Morales, Marco and Amato, Nancy M},
journal={arXiv preprint arXiv:2206.03467},
year={2022}
}Abstract
Evaluating Guiding Spaces for Motion Planning, Amnon Attali, Stav Ashur, Isaac Burton Love, Courtney McBeth, James Motes, Diane Uwacu, Marco Morales, Nancy M. Amato, IROS 2022, Workshop for Evaluating Motion Planning Performance, Kyoto, Japan, Oct 2022. @misc{https://doi.org/10.48550/arxiv.2210.08640, Randomized sampling based algorithms are widely used in robot motion planning due to the problem's intractability, and are experimentally effective on a wide range of problem instances. Most variants do not sample uniformly at random, and instead bias their sampling using various heuristics for determining which samples will provide more information, or are more likely to participate in the final solution. In this work, we define the motion planning guiding space, which encapsulates many seemingly distinct prior works under the same framework. In addition, we suggest an information theoretic method to evaluate guided planning which places the focus on the quality of the resulting biased sampling. Finally, we analyze several motion planning algorithms in order to demonstrate the applicability of our definition and its evaluation.
Keywords: Algorithms, Guidance
Links : [ArXiv] BibTex
doi = {10.48550/ARXIV.2210.08640},
url = {https://arxiv.org/abs/2210.08640},
author = {Attali, Amnon and Ashur, Stav and Love, Isaac Burton and McBeth, Courtney and Motes, James and Uwacu, Diane and Morales, Marco and Amato, Nancy M.},
keywords = {Robotics (cs.RO), Artificial Intelligence (cs.AI), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Evaluating Guiding Spaces for Motion Planning},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}
Abstract